Total Survey Error in Biomeasure Data Collection

Total Survey Error in Biomeasure Data Collection | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Total Survey Error in Biomeasure Data Collection Alexandru Cernat, Brad Edwards, Colter Mitchell, Joseph W. Sakshaug This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7711723/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Background The integration of biomeasures into social surveys has grown rapidly, enabling new insights into biosocial processes by combining probability-based survey data with biological indicators. However, biomeasure collection outside clinical settings introduces unique sources of error that are not fully addressed by traditional survey methodology frameworks. This paper adapts the Total Survey Error (TSE) framework to account for the distinctive challenges of biomeasure data collection in surveys. Methods We review the stages of biomeasure data collection in social surveys and systematically map them onto the TSE framework. We extend the framework to identify new error sources, including biomeasure nonresponse, data loss during transmission, laboratory and batch effects, and interviewer-related biases. The discussion draws on evidence from large-scale biosocial surveys and methodological studies, highlighting trade-offs between different sources of error and the implications for study design, data processing, and analysis. Results Our extended framework demonstrates that biomeasure collection adds multiple new opportunities for error across both selection and measurement processes. On the selection side, additional nonresponse mechanisms—such as refusal to consent to biological collection or ineligibility due to physical characteristics—can create systematic biases. On the measurement side, issues such as interviewer variability, shipment and storage conditions, laboratory practices, and batch effects can significantly influence data quality. Trade-offs emerge between maximizing measurement precision and minimizing participation bias, between standardization in clinical contexts and ecological validity in household settings, and between cost constraints and optimal data integrity. Conclusions The adapted TSE framework provides a structured approach to evaluating and minimizing errors in biomeasure data collection in social surveys. Recognizing and addressing new sources of error is essential for producing valid and generalizable biosocial research. Future work should focus on developing standardized protocols, improving interviewer and nurse training, expanding feasible self-collection methods, and quantifying the relative magnitude of different error sources. By systematically incorporating biomeasures into probability-based surveys while accounting for these challenges, researchers can strengthen the reliability and policy relevance of biosocial findings. Total Survey Error Biomeasures Biosocial surveys Data quality Measurement error Nonresponse bias Laboratory effects Survey methodology Figures Figure 1 1. Introduction In recent years, the integration of biomeasure data collection into social surveys (so-called “biosocial surveys”) has significantly expanded, leading to important new research that leverages the strengths of probability sampling from the general population and the precision of biological measurements. Nevertheless, combining social surveys with these measurements brings unique challenges regarding data quality. The Total Survey Error (TSE) framework has been traditionally used by survey researchers to guide decision-making processes regarding data collection and the evaluation of data quality ( 1 ). However, the increased complexity and the new actors involved in collecting bio-social survey data necessitate adapting this framework. Biomeasure collection, encompassing measures from anthropometrics to molecular biomarkers, presents new sources of potential error not typically encountered in traditional social surveys. These errors arise from the complexity of collecting biomeasures outside of controlled clinical environments, involving multiple actors such as nurses and lay interviewers. Each stage of this process, from initial sampling to post-survey adjustments, introduces opportunities for error, thus complicating the trade-offs necessary to maximise data quality. Due to the complex nature of these errors, adapting the TSE framework to encompass biomeasure collection is crucial. Traditional TSE components like noncoverage, nonresponse, measurement, and processing errors must be re-evaluated and expanded to include the challenges of collecting biomeasures. For instance, the potential for nonresponse (or nonconsent) bias increases when participants are reluctant to provide invasive biological samples or are too burdened to participate in any physical data collection. Further, implementing biomeasure collection outside clinical settings is inherently complex and error-prone. Logistics such as equipment provision, training, specimen collection, transportation, and processing require rigorous protocols to ensure high-quality data collection. These steps, including the involvement of various data collection personnel with differing levels of expertise and in different environments, can introduce variability, impacting the reliability and validity of the collected data. This article expands on the TSE framework to include the complexities of collecting biomeasures in social surveys, highlighting their unique challenges. By examining the existing components of the TSE framework as well as proposing new stages and types of errors, we provide a guide for researchers collecting biomeasures as part of social survey data collection. 2. Overview 2.1 Typology of Biomeasures The integration of biomeasures into social and health research, from anthropomorphic measures to glucose to brain structure to molecular measures, has allowed researchers to examine physical and biological mechanisms, precursors to disease, residues of exposure, and moderators of response ( 2 , 3 ). Scholars who study human health and development have been trying to integrate biological theory, methods, and data for centuries ( 4 – 8 ). And since the late 20th century, there has been a rapid integration of biological measures into social and health research, in part due to the rapid technological advancements that afforded these measurements ( 2 , 6 , 9 – 11 ). Several rationales have been provided for integrating bio-measures with social and behavioural data ( 5 , 7 , 12 , 13 ). A broad definition of a biomeasure is an objectively measured indicator of an organism, its functioning, or its systems ( 14 ). Because biomeasures have such a wide range of potential uses, they can often be challenging to classify. For example, for some researchers, biology is expected to be a mechanism linking exposures to health (i.e., the way for the environment to “get under the skin”), but they are also of interest as indicators of physical health status, exposure and disease, so that, even if they are not part of the causal pathway from external stimuli to outcome, researchers can reconstruct the past and potentially predict the future ( 8 , 10 , 12 , 15 ). Indeed, each system could have potentially hundreds of biomeasures, some of which are more typically used as causal mechanisms and others as predictive indicators. Similarly, biomeasures could be classified by how they are collected, in what tissue they reside (if any), or what function they measure. Most biomeasures were not developed with the goal of examining variation related to social behaviour, to study mechanisms linking the environment to child development outcomes, or even to be measured outside of a lab or clinic. Furthermore, most were developed on homogeneous clinical populations that are typically higher income, western, white, and European ( 5 , 8 , 10 , 15 ). As such, their use in social surveys needs to account for new types of errors that might arise. Although far from exhaustive, by looking at a number of large population-based studies (e.g., National Study of Adolescent to Adult Health; Future of Families and Child Well Being; Health and Retirement Study), biomeasure networks (e.g., Early Genetics and Lifecourse Epidemiology – EaGLE; Environmental Influences on Child Health Outcomes; NIA Biomarker Network), and biomeasure reviews ( 7 , 10 ), we discuss key biomeasures likely to be included in survey-based research studies (Table 1 ). Table 1 Main types of biomeasures and their data collection characteristics Biomeasure Collection mode Collection cost 1 Low − 5 High Susceptibility to shipping effects 1 Low − 5 High Common health measures analyzed Comment Key Survey Methods Citations Saliva Interviewer or self 2–3 2 DNA, RNA, microbiome, hormones, some drug use Well documented costs and issues Adam & Kumari, 2009; Zhang, Xiao, & Wong, 2009; Gavrilova & Lindau, 2009; Kaczor-Urbanowicz et al., 2017 Blood Venous Phlebotomist or nurse 4 5 Largest potential number of tests: DNA, RNA, hormones, proteomics, metabolomics, lipids, CBC, blood chemistry, inflammation, organ function, cholesterol, glucose Expensive to collect and ship, timing and temperature sensitive, most versatile post collection Giavarina & Lippi, 2017; Crimmins, Vasunilashorn, Kim, & Alley, 2008; Dagur & McCoy, 2015 Dried blood spots Interviewer or self 2 3 Reduced number, but similar types as venous Difficult to self-collect McDade, Williams, & Snodgrass, 2007; Ostler, Porter, & Buxton, 2014; Jacobson et al., 2023; El-Sabawi et al., 2024; Cella et al., 2010 Hair Interviewer or self 2 1 Hormones, toxicant exposure, drug use Not all respondents are eligible if hair is treated or too short Lawes, Hetschko, Sakshaug, & Eid, 2024; Ford, Boch, & McCarthy, 2016; Pragst & Balikova, 2006 Physiological functioning and anthropometric Interviewer 2–4 1 Gait speed, grip strength, pulse rate, temperature, blood pressure, lung capacity, sit-stand test, vision test, hearing test, smell test Equipment can be costly or inexpensive, depending on the specific measure Ulijaszek & Kerr, 1999; Casadei & Kiel, 2019; Gross, Jones, & Inouye, 2015 Stool Self 2–3 4 Microbiome Collections range from simple wipe collections to larger samples that allow for greater flexibility in use later. Loftfield et al., 2016; Sinha et al., 2016; Li et al., 2023 Urine Self 2 3 Drugs, toxicant exposure, hormones Fairly easy to collect, but many labs are not interested in holding John et al., 2016; Fenton, Johnson, McManus, & Erens, 2001 Task-based cognition Interviewer or self (via app) 1–2 1 Cognition, memory Can be difficult to administer Chalhoub-Deville, 2013; Hollnagel, 2003; Suzuki et al., 2004 Each biomeasure has strengths and weaknesses, some of which are summarized in Table 1 . Sources of measurement error are specific to the biomeasure collection method and play a significant role in the quality of the final estimates. In the table, we provide eight common biomeasures and discuss four broad categories of issues: 1) who collects the biomeasure, 2) relative collection cost, 3) potential issues, and 4) common health measures analyzed (though far from an exhaustive list). Venous blood collection provides a comprehensive sample that is suitable for the widest range of analyses. However, collecting venous blood in a home setting often requires a healthcare professional due to the invasiveness of the procedure, which can increase costs and reduce the convenience factor. And often, the initial processing in the field and cold storage shipping can be expensive if it is far from the processing lab. Dried blood spots (DBS) offer a practical alternative for home collection of blood samples, as they are stable at room temperature and can be easily mailed. The downside is that the volume of blood obtained is limited, which may affect the breadth of assays that can be performed, and extraction from DBS can sometimes be less efficient than from whole blood. And there are some known biases resulting from temperature and humidity exposures. Also, although DBS could be collected by the individual, evidence suggests this is challenging for most participants–with often low-quality samples. Newer devices that try to bridge the gap between venous blood and DBS, such as the TASSO device, are being tested more regularly now. Another minimally invasive mode of collection is saliva collection. It does not require specialized equipment and is well-suited for hormone (e.g., cortisol) and genomic studies in particular. Because it is simple to collect and non-invasive, it is ideal for use in large-scale self-collection efforts and in pediatric populations. However, the reliability of saliva samples can be influenced by recent eating or drinking and contamination with food particles or blood. And some hormones require time-specificity in the collection. Anthropometric measures, such as weight, height, and waist circumference, can be self-collected and serve as indicators of nutritional status or disease risk, but the accuracy of self-reported measurements can be compromised by a lack of precision or user error. Many of the other physiological assessments must be performed by a trained interviewer. Materials needed for these assessments can range from relatively inexpensive (a scale, tape measure, or stopwatch for each interviewer) to far more expensive devices, depending on the biomarker of interest. Hair samples provide a long-term biomarker analysis option, reflecting exposure to substances like heavy metals or drugs over months ( 16 ). They are relatively easy to collect (by an interviewer), ship, and store without the need for special conditions. However, external contamination (from treatments, dyes) can influence some measures, and participants with very short hair or semi-permanent procedures, or sensitivities around their hair, often must be excluded. Collecting stool samples at home is advantageous for microbiome studies, detecting gastrointestinal conditions, and parasite testing. The process is non-invasive, and samples can be mailed to the laboratory for analysis. However, the principal challenge is ensuring proper sample preservation and avoiding degradation, which can affect bacterial viability and microbial community analysis. Moreover, the collection process might be uncomfortable or unappealing for some participants, potentially impacting participation rates. Urine collection, on the other hand, is more straightforward, non-invasive, and ideal for detecting metabolites, hormones, and drug substances. Yet, variability in concentration due to hydration levels and the potential for sample contamination are recognized limitations. Also, the shipping can be particularly challenging if an interviewer isn’t taking it with them. Moreover, as with stool samples, fewer labs are equipped to work with these samples. Task-based cognition tests have been developed in clinical settings to assess a wide range of brain functions (memory, attention, and processing speeds, executive function, language and verbal fluency, and visuospatial ability) in various fields, including psychology, neurology, and gerontology ( 17 ). Many have been applied in surveys, especially in child studies and surveys of elderly populations. Structured tasks or activities are typical administration forms. Some task-based cognition tests can be self-administered, but others require interviewer administration to ensure standard conditions are met (such as adequate lighting and distance for a vision test) or to deploy specialized, costly devices (such as equipment for a smell test). 2.2 Models for Biomeasure Data Collection in Survey Research A range of models have been implemented for combining biomeasures and survey data collections ( 18 ). In the study design phase, new studies often have a wider range of potential biomeasure models to consider than studies in a later phase seeking to add biomeasure components to existing surveys (such as a long-running panel study). Tradeoff dimensions include administration mode, data collection personnel, collection location, processing location and procedures, and timing of consent and collection. Some studies may be created with biomeasure collection as a primary focus, and with survey collection as secondary or supplemental. The National Health and Examination Survey (NHANES) in the U.S. is one example. Much more common are established surveys (repeated cross-sectional or longitudinal) that add biomeasure components midstream (such as the UK Household Longitudinal Study or the English Longitudinal Study of Ageing). In either case, tradeoffs at each stage have cost and quality considerations, and extending the TSE framework can be helpful in thinking about biomarker collection and processing. Beebe ( 19 ) suggests three models for biomeasure collection in surveys: self-administered, interviewer-administered in the home, and administered by an interviewer or medically trained person in a clinic. Simpler biomeasure data (e.g., hair, saliva, urine, accelerometry) can be collected by participants themselves ( 20 – 22 ). Other biomeasures (e.g., blood, grip strength) require collection by trained staff in-person, often with special equipment or tools. Height and weight are interesting examples because self-reports are sometimes deemed too unreliable to include in a data set on their own, and are obtained by direct measurement in person. Skill requirements differ – field interviewers can be trained to collect some physical and cognitive functioning data ( 23 ), but others (e.g., venous blood) require a phlebotomist or nurse ( 24 ). If tests require highly skilled personnel (e.g, physicians, radiologists), the home may not be an appropriate setting, and travel to a clinic or other well-equipped location may be the only choice. Some studies have considered mobile vans as a compromise, with NHANES being a notable example ( 25 ). Bio-specimens may need to be shipped to a laboratory for processing under controlled conditions with specialized equipment and kept at controlled temperatures. Timing in relation to the data collection has two dimensions: time from the interview, and repetition across interviews on longitudinal or repeated cross-sectional surveys. Some biomeasure data is so tightly linked to the survey that it must be collected at the same time or very close to it. For example, the Population Assessment of Tobacco and Health (PATH) Study collects data about tobacco and drug use in an interview, and schedules a blood draw within 10 days to compare values from the blood with the self-reports, and to augment survey reports on health behaviors ( 24 ). Other data that are more stable over time may be collected in a looser relationship to the interview, for example, altimetry data measuring activity levels for a week. The other dimension is about repeating measures over time. For longitudinal and repeated cross-sectional surveys, it is often critical to collect the same measure repeatedly under the same conditions. Controlling the administration and processing of biomeasure data takes on even more importance in such comparative work. Combining biomeasure collection with survey collection can add additional error sources at every point in the process ( 19 ) due to use of nurse or phlebotomist, collection equipment, shipment, and laboratory. Each activity in the flow from design, training, and collection through post-collection processing adds friction that can lead to error. 2.3 Adapted TSE for Biomeasure Data Collection To better understand the implications of collecting biomeasure data in surveys, we propose to expand the TSE framework. Figure 1 builds on the traditional TSE framework while highlighting the new stages as well as their related errors. On the selection side (right-hand side of Fig. 1 ), in addition to the typical survey selection stages (such as target population, sampling frame, and sample), there are further stages for biomeasure respondents. This group can be significantly different from regular survey respondents due to eligibility criteria, consent requirements, and respondents’ abilities and motivation. Moreover, some biomeasures may need to be shipped to a lab or research center for processing or analysis. This stage can also lead to loss of data and selection, for example, due to items packed in dry ice but delivered too late to maintain the proper temperature. On the measurement side (left-hand side of Fig. 1 ), once the biomeasures are collected, the handling and shipment of them can impact their integrity and the quality of the measurements. This is especially the case for biological samples (e.g., tissue, blood), which may need to be sent to specialized labs. If the shipments take too long or if the packaging is not adequate, it can have an impact on the measurements. Additionally, once the specimens arrive, the lab and the batches in which they are analysed can influence subsequent results. In the next sections, we discuss each stage of the revised TSE diagram and highlight how these stages and errors can affect the collection and quality of biomeasure collections. 3. TSE Stages in Biomeasure Research: Selection This section discusses biomeasure error in the selection process, from the target population to the sampling frame, from the frame to sample selection, from sample selection to respondents, from respondents to raw data, and from raw data to post-survey adjustment. For each of these stages, we discuss issues that are specific to surveys with biomeasure data collection. 3.1 Target Population to Sampling Frame: Coverage Error The first stage of selection identifies the population of interest and the sampling frame that will be used. The difference between these two leads to coverage error. The size of the error depends on the proportion of people missing from the sampling frame and how different they are from the general population in terms of the variables of interest. If the survey design starts from a probability sample of the general population, followed by biomeasure data collection, then coverage error is similar to that present in social surveys. Additional errors could appear if there is a relationship between the likelihood of people being in the sampling frame and the biomeasures of interest. For example, people with poorer health, and thus different levels of biomeasures, might be less likely to have a phone (and not be covered by a random-digit-dial sampling procedure) or registered to vote (if voting records are used for sampling). If the data collection is based on a non-probability selection process, such as the UK Biobank or the US All of Us Research Program, some people will have no chance of being selected. In this case, it is difficult to separate coverage error from non-response error (see next section), leading to selection bias. This can be especially problematic if the process underrepresents important groups such as ethnic minorities. Researchers also need to consider whether the definition of the target population needs to be adjusted due to the collection of biomeasures. This is especially the case if some biomeasure sub-groups have no chance of being selected. For example, cortisol from hair samples cannot be collected from men who have very short hair or are bald. In this case, either the target population is defined in such a way as to exclude such people or alternative measures need to be considered (e.g., using saliva or urine). 3.2 Sampling Frame to Sample: Sampling Error The next stage in data collection is using a sampling frame (a list of the population of interest) or a procedure to select households or individuals. If the selection process is random, it can lead to a degree of uncertainty, which is known as sampling error. The size of the sample and the procedure used can impact the amount of sampling error. When collecting biomeasure data as part of a survey, the size of the sampling error is not different unless the procedure is influenced by the biomeasure collection. This could happen if, for example, biomeasure information is collected under a stratified or clustered sample design. In studies where cases are selected using a non-probability selection process, there is often a tradeoff between sampling error and systematic bias due to coverage and non-response. For example, a large study such as the UK Biobank will have apparently precise estimates due to the large sample size. Nevertheless, the estimates may be biased, with systematic selection biases due to coverage and non-response errors. One of the reasons for the increasing popularity of combining probability samples with biomeasure data collection is to avoid the perils of such biased samples. 3.3 Sample to Survey Respondents: Survey Nonresponse The next stage in the selection process is collecting data from all sampled cases. When some individuals decide not to participate in the study, there can be a nonresponse error. As in the case of coverage error, the amount of error depends on the proportion of cases that do not participate and how different they are from the participants. There are numerous mechanisms that can affect the likelihood of participating in a survey that involves biomeasure data collection. While some individuals may be averse to participating in such a study, others may find the study’s focus to be interesting and salient to their lives, which may encourage them to participate. Research has shown that the study’s topic plays an important role in influencing survey participation, in addition to other survey design features, such as the offering of incentives and data collection mode ( 26 ). Biomeasure data collection can also have detrimental effects on retention in longitudinal studies because of the extra respondent burden. In a quasi-experimental study in the UK, Pashazadeh et al. ( 27 ) exploited a scenario where certain participants were randomly selected for a nurse visit for biological data collection. They found a short-term negative effect on subsequent wave cooperation (wave 3) for those included in the nurse visit, suggesting that the increased burden led to a slight decrease in participation rates following the visit. However, this effect was not observed in later waves ( 4 – 7 ), indicating that the initial negative impact of the nurse visit on survey cooperation did not persist over time. 3.4 Survey Respondents to Biomeasure Respondents: Biomeasure Nonresponse In addition to the nonresponse mechanisms present in social surveys, collecting biomeasure data can introduce additional sources of nonresponse error. Firstly, some studies collect biomeasures at multiple stages (e.g., a survey followed by a separate biomeasure data collection interview) undertaken by different actors, e.g., nurses. This could lead to additional nonresponse. Secondly, biomarkers often need specific consent before the measure can be collected, which can be either verbal or written. Consent requests can increase the amount of nonresponse. Third, the collection of some biomeasures might not be possible, especially for certain respondents. For example, walking speed cannot be collected for non-ambulatory participants; hair cannot be collected if participants are bald or have dyed hair; it might be impossible to collect full blood from some participants; and a vision test cannot be conducted with people in advanced stages of Alzheimer’s disease. The mechanism for the nonresponse in these cases depends both on the respondent's characteristics and on the interviewer/nurse’s skill. Finally, for many studies, respondents need to agree to allow the biomeasure data or results derived from it to be archived and shared with other researchers for future analyses. There is some evidence regarding nonresponse that is specific to biomeasure data collection. For example, willingness to participate in biomeasure data collection varies considerably by topic. Boyle and colleagues ( 28 ) reported that while about 77% of respondents showed willingness to participate in a less burdensome survey on an interesting topic, only 62%-64% expressed willingness to participate in more invasive health surveys involving measurements like blood pressure or waist circumference, and just 45% were open to a finger stick blood draw, highlighting potential nonresponse biases in surveys with biomeasure due to their intrusiveness and the burden they impose. Among those who completed the Adult Interview in an actual biomeasure collection study, the weighted response rates 63.6% for providing a urine sample and 43.0% for providing a blood sample, providing further evidence of an intrusiveness effect. In another study, Sakshaug et al. ( 29 ) examined factors influencing consent to physical measurements in the Health and Retirement Study, focusing on older adults. The study found that while 79% of respondents consented to all physical measurements, factors such as demographics, health status, levels of survey cooperation, and interviewer characteristics significantly influenced consent rates. Key findings indicated that Hispanics paired with interviewers bilingual in Spanish and English, people with recent doctor visits, and people with diabetes were more likely to consent, whereas younger individuals, people with functional limitations, or people who were reluctant survey participants were less likely to consent to biomeasure collection. There is also some evidence regarding biomeasure nonresponse patterns in general population surveys. For example, in large panel studies in the UK, older individuals, those with higher educational attainment, homeowners, and white respondents generally showed higher participation in biomeasure collection ( 30 ). Conversely, women were less likely to provide blood samples despite their higher engagement in initial stages like nurse visits. Household dynamics, such as living alone and having a partner, can affect participation across different stages. Additionally, both long-term illness and good to excellent self-reported health positively influenced participation, underscoring how personal health status and socioeconomic factors shape engagement in such studies. The National Social Life, Health, and Aging Project’s (NSHAP) implementation of in-home salivary specimen collection from older adults reported that 90.6% of participants provided saliva samples, with a significantly lower participation rate among Black participants compared to other racial groups ( 31 ). Interestingly, factors such as age, gender, education, and self-rated health did not significantly influence the likelihood of providing a saliva sample, highlighting specific demographic patterns in the participation of older adults in biomeasure data collection within large social surveys. Impact of data collection design on participation There is some evidence for differences in participation based on the mode of the biomeasure collection. An experimental study in the UK showed that self-completion of dried blood spots had significantly lower participation compared to that done by a nurse (35% vs. 74%), although there were no differences in participant characteristics ( 32 ). There is also limited research on the impact of offering feedback as a way to encourage participation in biomeasure collection. Benzeval et al. ( 33 ) have shown that there is no effect of offering feedback on the likelihood of collecting a blood sample in a panel study in the UK, but it did have an effect on consent to giving blood. There was also some evidence of differential effects with the highest impact on web participants. While design decisions such as mode and giving feedback aim mainly to maximise participation and minimise costs, they may also have unintended consequences on the measurement and behaviour. For example, receiving feedback about their health could encourage people to exercise or change their eating habits. From a research perspective, this is a particular concern in longitudinal studies and could lead to a kind of conditioning effect unique to biomeasure collection. Interviewer/nurse impact on non-response Nonresponse can also be influenced by the interviewer's skills and characteristics, as highlighted by West and Bloom ( 34 ). This can also be the case for biomeasure collection. Looking at two waves of the UK Household Longitudinal Study and ELSA, researchers found that nurses accounted for between 5% and 14% of the variance in nonresponse rates to biomeasure collection. The impact of the nurses varied by stage of data collection, with 5%, 9%, and 14% of the variation explained for nurse visits, consent to blood, and blood collection. There was some evidence of effects of nurse characteristics, with older nurses generally being more successful at obtaining cooperation and consent for data collection, while nurses with more survey experience were more likely to successfully collect blood samples ( 30 ). For longitudinal studies with multiple waves of biomeasure data collection, previous research shows that switching nurses between survey waves does not significantly impact respondents' participation in biomeasure collection. This includes participation in the nurse visit, consent to blood collection, and successful blood collection ( 35 ). Furthermore, in some cases, it can be beneficial, especially for respondents who did not participate in the previous wave's biomeasure collection. This suggests that nurse continuity is not critical for ensuring high participation rates in subsequent waves of biomeasure collection. 3.5 Biomeasure Respondents to Raw Data: Data Loss Errors The process of converting a biological specimen to usable data is complex and has several steps: transfer from the field to the lab, inventorying an intermediate or analytic lab, processing, assay analyses, and quality control assessments. Each step provides an opportunity to lose, contaminate, or modify samples. However, most research on lab quality focuses on issues of labs influencing measurement error rather than representation. There are at least two major areas of concern for representation. First, loss of sample, which reduces statistical power, and if that loss is correlated with a variable of interest, bias can result. Because of the cost of obtaining the sample in the first place, the price of losing the sample is high. For that reason, most studies employ detailed shipping and storage protocols. And many labs will rerun samples that fail quality control so as to obtain as much information as possible. Relatedly, the number of times a sample is transferred to another lab or location, the more likely it is to be lost or unintentionally modified. A second, more likely scenario, is that initial protocols are modified to fix discovered flaws, and if those early participants are more likely to be from certain populations (i.e., more compliant, the first geographic target, etc), then there could be a bias in who will provide a viable sample. 3.6 Postsurvey Adjustment: Adjustment error As the previous sections have highlighted, the mechanisms of nonresponse in biomeasure data collection can be different from those of participating in a social survey and answering questions. As such, the strategies for correcting for nonresponse in biomeasure data should account for that. One strategy is to separate each stage of nonresponse and create conditional models that explain the causes of missingness for each stage. For instance, Cernat et al. ( 35 ) developed regression models to analyze participation in nurse visits, consent to collect blood data, and the actual blood data collection. Their analysis revealed distinct predictors for each mechanism. When generating multiple weights to adjust for each mechanism, they observed no significant differences in the impact on point estimates for pulse and C-reactive protein (CRP). However, incorporating adjustments for biomeasure data collection yielded estimations that differed from those derived solely from using survey nonresponse weights. Many large studies like the UK Household Longitudinal Study, English Longitudinal Study of Ageing, Health and Retirement Study, or the China Health and Retirement Longitudinal Study provide nonresponse weights that account for some of the selection processes in biomeasure data collection. Other models can also be used to correct for nonresponse in biological data collection. For example, Clarke and Houle ( 36 ) employed Heckman selection models to adjust for nonresponse biases in biomeasure data from HIV prevalence studies. This method was specifically used to correct for selection biases due to differential nonresponse in demographic and health surveys. The Heckman model provided a way to estimate the correlation between the likelihood of participating in HIV testing and the actual HIV status, thereby correcting the HIV prevalence estimates to better reflect the true population values. Different strategies for adjusting for nonresponse in biomeasures were explored in a simulation study using data from the English Longitudinal Study of Ageing ( 37 ). In it, researchers examined the effects of missing data on socio-economic analyses using the C-Reactive Protein biomarker. The study explored three missing data mechanisms—Missing Completely at Random (MCAR), Missing at Random (MAR), and Missing Not at Random (MNAR)—and tested four compensation methods: inverse propensity weighting, a Heckman selection model, multiple imputation, and a hybrid approach combining the Heckman model with multiple imputation. The findings indicated that multiple imputation is the most effective method for addressing missing data across all scenarios, making it a robust choice for biosocial research. Best practices regarding correction for nonresponse vary by discipline and journals. Some studies have no weights available to deal with nonresponse in biomeasure data. In such cases, users should consider creating their own or using alternative methods, such as multiple imputation. Even when weights do exist, some researchers may decide not to use them due to the extra complexity involved or uncertainties regarding best practices. 4. TSE Stages in Biomarker Research: Measurement We next discuss the second branch of the Total Survey Error framework: measurement. Here, we review how the process of going from the construct of interest to the raw data can be different when collecting biomeasures. We also introduce new types of errors: transmission and lab and batch effects, that can be especially important when collecting biological data. 4.1 Construct to Measurement: Validity The extent to which a biomeasurement is valid, or captures what the researchers intend, is highly dependent on the concept of interest. Some biomeasures have been extensively tested, while others are more novel and have had limited validation. Typically, the focus is either on validating using a gold standard clinical measure (e.g, interview measured weight vs. a physician measured weight) or using the predictive capacity of a health or behavioural outcome (clinical or predictive validity ( 38 )). When a clinical comparison is unavailable, criterion validity is commonly used by examining whether the biomeasure correlates with known covariates. Typically, significant effort is made during the assay or protocol development to examine the validity of the biomeasure, but with more novel measures, or those biomarkers with a large number of potential algorithms (e.g., genetics), few validity analyses have been conducted. One tension for survey-paired biomarkers is that biomeasures are typically compared to a parallel test conducted in an ideal context (i.e., a clinic) vs. in-home, yet often the purpose of many studies is to provide more naturalistic measures. So, a critical question is whether an observed deviation from the gold standard is evidence of lower validity or potentially a more real-life measurement. The degree to which contextual effects lead to a systematic change in the measurement can lead to a lack of validity. Lack of validity of biomeasures can substantially undermine their utility in research contexts, leading to misleading scientific conclusions. One major issue is the over-reliance on biomeasures without adequate validation. Many biomeasures are reported based on initial discovery studies but lack subsequent rigorous validation in independent, representative cohorts. Such premature reliance can result in false assumptions about their predictive or diagnostic power. Another prevalent misuse concerns the lack of consideration for confounding variables and proper standardization across studies. Biomeasurements such as C-reactive protein (CRP) and interleukins can be influenced by a variety of non-disease factors, including age, sex, diet, and circadian rhythms. Without adjusting for these confounding factors, studies may draw erroneous associations between biomeasures and disease states. The lack of validity often stems from commercial pressures and the premature commercialization of biomarker assays. While the commercialization of biomeasure tests has brought innovative diagnostic tools to research, it has also led to the widespread availability of tests that may not have undergone thorough validation. For instance, genomic biomarkers such as polygenic scores, epigenetic clocks, and telomere length are often put on the market based on limited evidence, which can result in their misuse. 4.2 Measurement to Response: Measurement Error Once the measurement instruments are developed to capture a concept of interest, they have to be applied to collect a response from study participants. If the instrument does not capture the intended measure, this results in a measurement error. In the case of biomeasures, measurement error can occur in several ways. Instruments might have been developed and pre-tested in ideal contexts and environmental factors, such as the biomeasure collection location and timing, that can lead to some degree of measurement error ( 39 – 41 ). Measurement error could also arise from fine-tuning instruments and clinical cut-offs that were developed in ideal conditions but do not meet in the real world, or that were based on specific populations (e.g., white males) that may not be appropriate for generalizing to more heterogeneous populations ( 42 ). Interviewer/nurse effects on biomarkers Another cause of measurement errors could be due to the interviewers. For example, a large study looking at eight waves of Survey of Health, Ageing and Retirement in Europe (SHARE) revealed moderate-to-large interviewer effects on physical performance measures in a cross-national survey, with variance partition coefficients (VPCs) ranging from 0.05 to 0.28 for biomeasures like timed chair stand, walking speed, grip strength, and peak flow, in stark contrast to the minimal effects observed for self-reported measures like height and weight ( 43 ). Notably, peak flow demonstrated the highest susceptibility to interviewer effects, followed by chair stand, walking speed, and grip strength. Additionally, significant cross-national differences in interviewer effects were identified, with particular biomeasures showing unexpectedly high VPCs in certain countries, aligning with cross-national variations found in other survey measures. The analysis also highlighted the absence of a consistent trend in interviewer effects across data collection waves, suggesting that increased interviewer experience or training over time does not uniformly improve data quality for all biomeasures, and in some cases, might even introduce variability in measurement errors. A similar study based on data from two UK surveys, the English Longitudinal Survey of Ageing and the UK Household Longitudinal Study, assessed how nurses influence the measurement of fourteen different health measures, including weight, height, pulse, grip strength, and lung capacity ( 44 ). They found a medium overall effect of nurses on measurement, with nurses accounting for around 13% of the variance in all measures combined. The effect varies significantly among different measures, ranging from approximately 2% to 25%. Specifically, grip strength and lung capacity measurements are more substantially influenced by nurses compared to other measures like height, weight, and pulse. Furthermore, the study reveals that characteristics of nurses only explain a very small proportion of the observed variance, suggesting that more research is needed to understand the mechanisms behind nurse-related measurement errors. A different study focusing on interviewer biological data collection in the US study National Social Life, Health, and Aging Project (NSHAP) showed large variation of the interviewers on biological measures ( 45 ). Touch sensory tests exhibit especially high interviewer effects, with about 30% of variation attributable to interviewers, while smell tests and timed balance/walk/chair stands show moderate interviewer effects around 10%. The presence of third parties or the type of dwelling did not explain this observed variation. 4.3 Response to Specimen: Transmission Error Shipping conditions can also introduce significant measurement errors in the assessment of common biological samples by altering their integrity. Variations in temperature and delays during the transportation of blood and serum samples can lead to the degradation of sensitive biomarkers such as cytokines and hormones. The lack of precise temperature control can cause hemolysis, proteolysis, and other biochemical changes that skew biomarker concentrations. These changes may result in false-negative or false-positive results, complicating clinical diagnostics and patient management. The handling and packaging protocols during shipping also play a crucial role in reducing measurement errors. Improper handling methods, such as inadequate insulation or poor packaging materials, can exacerbate the degradation of samples. Steps like using dry ice for freezing samples, employing temperature loggers to monitor conditions, and ensuring rapid transit times are critical for preserving the integrity of biomarkers such as lipids and proteins. Consistently following standardized shipping protocols can thus minimize pre-analytical variability and ensure more accurate and reliable biomarker measurements. Moreover, the duration of storage and transit has a direct impact on the stability of biomeasurements. Extended shipping times, especially under suboptimal cooling, can lead to the denaturation or degradation of biomarkers like insulin and vitamins, affecting their measurable concentrations ( 46 – 49 ). These studies underscore the importance of minimizing the time between sample collection and processing to mitigate these effects. Utilizing expedited shipping services and well-maintained cold chain logistics can significantly reduce measurement errors associated with shipping, thereby enhancing the reliability of biomeasure data used in clinical and research settings. 4.4 Specimen to Raw Data: Lab and Batch Effects Laboratory conditions and procedural variability significantly influence the measurement error of common biomarkers such as glucose, cholesterol, inflammatory markers, molecular measures, and hormones ( 6 , 50 , 51 ), even after accounting for pre-lab variables, including sample collection and handling. Studies highlight that different analytical techniques, reagent lots, and calibration procedures can yield varying results for the same biomarker. For instance, the accuracy and precision of enzyme-linked immunosorbent assay (ELISA) kits can differ between manufacturers, leading to inconsistent detection levels of biomarkers like C-reactive protein (CRP) and telomere length. These variations underscore the critical need for standardization of analytical methods and the implementation of rigorous quality control protocols to reduce measurement errors. Inter-laboratory variability also remains a significant challenge. Studies have shown that even standardized methods can produce different results when conducted across different laboratories due to differences in equipment calibration, technician expertise, and environmental conditions ( 52 , 53 ). For example, when testing for glycosylated hemoglobin (HbA1c) and cholesterol, differences in laboratory practices can lead to significant divergence in results, potentially impacting the diagnosis ( 54 ). Therefore, harmonization of methods and adherence to stringent inter-laboratory proficiency testing are essential measures to minimize errors and ensure reliable biomeasurement across diverse clinical settings ( 55 ). For many biomeasurements, harmonisation of standards and protocols has not yet occurred–particularly for field-collected studies. A related potential bias comes from batch effects. Batch effects refer to systematic technical variations that occur when samples are processed in different batches (same lab but at a different time–typically a few days or weeks separation), and these can significantly impact the measurement of common bio-measures. Like many lab effects, batch effects can arise from differences in reagent lots, calibration standards, and instrument performance over time. These variations can introduce biases that obscure true biological signals, leading to inconsistent and unreliable biomarker readings. For example, in large epidemiological studies, blood samples might be collected over several months or years, and if processed in different batches, batch effects may distort the accurate biomeasurements such as lipid profiles or cytokines, ultimately confounding the study results. The influence of batch effects is particularly pronounced in high-throughput omics technologies where large datasets are generated ( 56 – 60 ). In epigenetics, proteomics, and metabolomics studies, batch effects can drive significant discrepancies in biomarker quantification. These systematic differences can obscure biological variation and lead to erroneous conclusions. Strategies to mitigate these influences include randomized sample processing, the use of quality control samples across all batches, and the application of advanced statistical methods to adjust for batch variability. Properly accounting for batch effects is crucial to ensure robust and reproducible quantitative biomeasure analysis. Without rigorous quality control and validation, batch effects are difficult to detect For biomeasures that require long-term storage in the freezer, the freeze-thaw process can lead to significant measurement errors in the assessment of common bio-measures by causing physical and chemical alterations to biological samples ( 61 – 64 ). Repeated freeze-thaw cycles can lead to the degradation of proteins, nucleic acids, and other biomolecules found in serum and plasma samples. This degradation results from mechanical stress and the formation of ice crystals, which can disrupt cell membranes and denature proteins. For instance, common biomarkers like cytokines and growth factors are particularly sensitive to these structural changes, leading to decreased concentrations and variability in measurement. Similarly, lipid components such as cholesterol and triglycerides can undergo hydrolysis and oxidation during freeze-thaw events, resulting in altered levels and compromised measurement accuracy. These changes are especially problematic for longitudinal studies where samples are stored and analyzed over extended periods. Another dimension of freeze-thaw cycle effects involves the stability of molecular biomarkers such as RNA and DNA. Nucleic acids are susceptible to degradation through repeated freezing and thawing, which can greatly affect the accuracy of PCR and sequencing-based biomarker assays ( 65 ). Ensuring minimal freeze-thaw cycles, along with optimized storage conditions, is critical to maintaining the fidelity of biomeasures, thus ensuring the reliability of clinical and research outcomes. The CLIA (Clinical Laboratory Improvement Amendments) program is specifically designed to set quality standards for laboratory testing conducted on human specimens, particularly for diagnostic purposes in healthcare settings across the U.S ( 66 , 67 ). To do this, laboratories must undergo a structured process to ensure they meet federal standards for clinical testing performed on human specimens. Based on the complexity of tests performed—categorized as waived, moderate complexity, or high complexity—the lab will receive a specific CLIA certification type ( 68 ). Laboratories performing moderate and high complexity tests must pay a fee, which varies according to the complexity level and the volume of tests performed. Labs performing moderate and high complexity testing, inspections are required to assess compliance with CLIA regulations. These labs must also participate in proficiency testing, which involves periodic external assessments to ensure the accuracy and reliability of their testing results. Waived test laboratories are generally not routinely inspected but must ensure they meet basic quality standards. While CLIA is a U.S.-based program, other countries have their regulatory frameworks and accreditation systems for clinical laboratories. For example, many countries recognize the ISO 15189 standard, which is an international accreditation for medical laboratories focusing on quality management and technical competence. 4.5 Raw Data to Edited Response: Processing Errors Quality Control (QC) metrics are crucial in minimizing measurement error in the assessment of common biomarkers by ensuring the reliability and reproducibility of laboratory results. QC procedures help in identifying and correcting systematic errors that may arise from instrument performance, reagent variability, and procedural inconsistencies. Through the regular evaluation of control samples with known concentrations, laboratories can monitor the accuracy and precision of the assays employed. For instance, in the measurement of glucose and cholesterol levels, QC metrics can detect and rectify drifts and shifts in the assay performance, thereby ensuring that subsequent measurements remain within acceptable error margins. The implementation of QC metrics not only helps in maintaining the consistency of test results within a single laboratory but also enhances comparability across different laboratories. However, sometimes even small differences in advanced statistical QC methodologies have profound implications on measurement error in biomarker analysis ( 69 – 71 ). For example, seemingly small, arbitrary differences in the quality control metrics for RNA can result in significantly different counts of transcripts–yet most studies only apply one QC protocol with little examination of their effect. Yet for other biomarkers, the QC metrics are rigid, which can sometimes exclude more samples from some populations. Therefore, the integration of robust QC metrics into laboratory practices is indispensable for maintaining the integrity and utility of bio-measure data, but only if the QC protocols have been thoroughly examined in multiple populations. Task-based cognitive tests are subject to different processing errors. In conventional parlance, “lab” signifies a brick-and-mortar place where physical specimens are received and processed. However, as detailed in Table 1 , biomeasures can take other forms – cognitive, physical, and mental functioning, sensory, etc, and are subject to their own processing errors ( 72 , 73 ). Output from a sensor device, such as a high-precision accelerometer, for example, captures a person’s movement in digital form, and that raw data must be edited to create data for analysis. The editing process may require skills or tools the survey team does not possess, and the raw data must be transmitted to another organization for editing. In this case, the potential for processing error is analogous to that for physical specimens. 5. Trade-offs Between Types of Errors TSE offers a framework for understanding the mechanisms that can lead to different types of errors. Investigating each data collection stage can help us better estimate and prevent potential issues that can decrease data quality. Another strength of the framework is highlighting the tradeoffs often involved in data collection and how different types of errors are connected. In biomeasure collection, one example of this is the tradeoff between sampling error (precision) and selection bias. Studies like the UK BioBank have large samples that can lead to an apparent increase in precision. Nevertheless, this comes at the cost of a likely selection bias, with significant groups being under-represented. Biomeasure collection in probability surveys focuses on minimising selection bias with a resulting loss in precision. Another tradeoff within survey biological data collection is between nonresponse and measurement. More detailed or precise measures, such as biomarkers measured from blood have, in principle, better measurement quality than alternative indicators but come with a decrease in sample size and potentially more selective samples. Collecting blood and extracting biomarkers from it entails explicit consent, blood draw, and complex processing of the blood. All of these can lead to a more selective sample. Yet another tradeoff that may appear between internal and external validity in choosing the location for specimen collection, physical or cognitive tests. Collecting data in a clinic ensures that the context and the machines used are standardised, increasing the measurement quality. Nevertheless, the artificial nature of the context can also decrease external validity. Similarly, collecting biomarkers in people's homes can be more realistic and align with their behaviour. Still, there may be more measurement errors due to differences in contextual factors and the lack of validation of instruments outside clinical contexts. Study designers and users of biosocial data must be aware of these tradeoffs. One strategy is to minimize the total error of the survey, focusing on the main analyses and questions that will be investigated. Unfortunately, estimating total error is often challenging without “gold standard” measures. An alternative approach is to minimise the errors most relevant to their research questions while accounting for possible unintended consequences on the other dimensions of the TSE. Researchers also need to account for costs (see Table 1 ). More intrusive biomeasures entail using staff with specialised skills and the costs of processing biological data collection. Additionally, collecting such data must account for the extra costs of processing and storing it. As a result, data collection is often a process of optimising data quality while keeping within budget constraints. 6. Conclusions This paper has explored the application of the Total Survey Error (TSE) framework to biomeasure data collection in social surveys, identifying the unique challenges and sources of error involved. The integration of biomeasures with survey data offers an invaluable new source of information, allowing researchers to investigate complex biosocial relationships. However, the process also introduces new error sources at various stages, from participant selection to data processing, necessitating an adaptation of the TSE framework. The updated TSE framework highlighted some new data collection stages and associated errors that researchers need to consider. On the measurement side, researchers need to be aware of the stage of specimen and raw data collection. These stages can lead to two new types of errors: transmission error and lab/batch effects. On the selection side, we also highlight two new steps: biomeasure response and raw data which can lead to two new types of errors: biomeasure nonresponse and data loss error. In reviewing the framework we also discussed how the other stages of the TSE framework can be influenced by biomeasure data collection. A crucial point highlighted in this paper is the importance of using probability samples to make biosocial research better reflect the general population. By integrating biomeasures into probability-based social surveys, researchers can obtain more generalizable findings, which are vital for public health and policy-making. Developing best practices and sharing knowledge across the research community are essential for advancing the field of biosocial surveys. Unfortunately, much of the practical knowledge and methodological innovation in this area remains unpublished. It is vital to document and disseminate these insights to establish standardised protocols and improve data quality across studies. Further research is also needed to understand the effects of biomeasure collection on nonresponse. As collecting biomeasures often requires participants to undergo invasive procedures, this can impact response rates and introduce biases. Balancing the improved measurement accuracy from biomeasures with potential increases in nonresponse bias is essential for making informed decisions about study design. As many traditionally interviewer-administered surveys are moving towards self-completion, there is also a need to study the feasibility of collecting a broader range of biomeasures in self-administered surveys. A key question is which biomeasures respondents are willing to collect themselves, and whether such measures can be collected with the necessary precision and accuracy without the aid of a professional. The impact of interviewer effects on the collection of biomeasure data is another area requiring more investigation. Interviewers or nurses can add significant variability in the collection of biomeasure data, and there is increasing evidence that this phenomenon is widespread. More research is needed to understand how data collection agents impact substantive analyses. Additionally, investing in better training programs and developing standardised protocols can help minimize these effects and improve data reliability. There is also a significant gap in our understanding of how biomeasure data collection protocols need to be adapted for use outside of clinical settings. Collecting biomeasure data in diverse environments such as homes or mobile sites poses unique challenges in maintaining data integrity and consistency. More research is needed to develop appropriate protocols and scaling methods for these non-traditional settings. Another area that needs further development is understanding the relative sizes of the errors discussed in this paper. The TSE framework encourages us to consider multiple types of errors and the tradeoffs that appear during data collection. To inform data collection decisions, we need a better understanding of the magnitude of different types of errors and their relative importance. Finally, the consistent use of terminology and frameworks like the TSE framework is crucial for accurately discussing and addressing the challenges in biomeasure data collection. Concepts such as target population and selection bias are fundamental for understanding data quality and should be uniformly applied across studies. The TSE framework provides a valuable tool for systematically addressing these issues, ensuring more valid, comparable results in biosocial research. Declarations Ethics approval and consent to participate Not applicable Consent for publication All authors have read and approved the manuscript and agree with its submission to BMC Medical Research Methodology. All authors meet authorship criteria, and all contributions and funding (if any) are disclosed in the manuscript. Funding Not applicable Author Contribution Conceptualization and methodology (adapted Total Survey Error framework): A.C., B.E., C.M., J.W.S. B.E. wrote the first draft of the Introduction and the section on models for biomeasures. C.M. wrote the first drafts of the typology of biomeasures and the sections on laboratory and batch effects, transmission error, and processing errors. A.C. drafted the remaining sections, integrated contributions across authors, and coordinated the manuscript. All authors (A.C., B.E., C.M., J.W.S.) edited and revised the manuscript, approved the final version, and agree to be accountable for all aspects of the work. A.C. is the corresponding author. Acknowledgement Not applicable Data Availability Not applicable References Groves RM, Lyberg L. Total Survey Error: Past, Present, and Future. Public Opin Q. 2010;74(5):849–79. Harris KM, Schorpp KM. Integrating biomarkers in social stratification and health research. Annu Rev Sociol. 2018;44(1):361–86. Lenzenweger MF. Thinking clearly about the endophenotype–intermediate phenotype–biomarker distinctions in developmental psychopathology research. Dev Psychopathol. 2013;25(4pt2):1347–57. Meloni M. How biology became social, and what it means for social theory. Sociol Rev. 2014;62(3):593–614. Meloni M. From boundary-work to boundary object: How biology left and re-entered the social sciences. Sociol Rev. 2016;64(1suppl):61–78. Mitchell C, Schneper L, Notterman D, Faul J. Collection and Storage Effects on DNA Methylation and Telomere Length. 2024. Qian W, Zhang C, Piersiak HA, Humphreys KL, Mitchell C. Biomarker adoption in developmental science: A data-driven modelling of trends from 90 biomarkers across 20 years. Infant Child Dev. 2024;33(1):e2366. Mitchell C, Schneper LM, Notterman DA. DNA methylation, early life environment, and health outcomes. Pediatr Res. 2016;79(1):212–9. Freese J, Li JCA, Wade LD. The potential relevances of biology to social inquiry. Annu Rev Sociol. 2003;29(1):233–56. Justice JN, Ferrucci L, Newman AB, Aroda VR, Bahnson JL, Divers J, et al. A framework for selection of blood-based biomarkers for geroscience-guided clinical trials. Geroscience. 2018;40:419–36. Kaplan JM, Wong HR. Biomarker discovery and development in pediatric critical care medicine. Pediatr Crit Care Med. 2011;12(2):165–73. Falk EB, Hyde LW, Mitchell C, Faul J, Gonzalez R, Heitzeg MM, et al. What is a representative brain? Neuroscience meets population science. Proc Natl Acad Sci. 2013;110(44):17615–22. Mitchell C, McLanahan S, Brooks-Gunn J, Garfinkel I, Hobcraft J, Notterman D. Genetic Differential Sensitivity to Social Environments: Implications for Research. Am J Public Health. 2013;103(S1):S102–10. Mayeux R. Biomarkers: Potential uses and limitations. NeuroRx. 2004;1:182–8. Singh I, Rose N. Biomarkers in psychiatry. Nature. 2009;460(7252):202–7. Pragst F, Balikova MA. State of the art in hair analysis for detection of drug and alcohol abuse. Clin Chim Acta. 2006;370(1–2):17–49. Hollnagel E, editor. Handbook of Cognitive Task Design. CRC; 2003. Sakshaug JW, Ofstedal MB, Guyer H, Beebe TJ. The Collection of Biospecimens in Health Surveys. In: P. Johnson T, editor. Health Survey Methods [Internet]. 1st edn Wiley; 2014 [cited 2025 Sept 25]. pp. 383–419. Available from: https://onlinelibrary.wiley.com/doi/ 10.1002/9781118594629.ch15 Beebe TJ. Challenges to Collecting Survey-Based Biomarker and Genetic Data. Surv Res. 2007;(3). Allicock OM, Petrone ME, Yolda-Carr D, Breban M, Walsh H, Watkins AE, et al. Evaluation of saliva self-collection devices for SARS-CoV-2 diagnostics. BMC Infect Dis. 2022;22(1):284. Nunes AP, Oliveira IO, Santos BR, Millech C, Silva LP, González DA, et al. Quality of DNA extracted from saliva samples collected with the Oragene ™ DNA self-collection kit. BMC Med Res Methodol. 2012;12:1–6. Zarbl J, Eimer E, Gigg C, Bendzuck G, Korinth M, Elling-Audersch C, et al. Remote self-collection of capillary blood using upper arm devices for autoantibody analysis in patients with immune-mediated inflammatory rheumatic diseases. RMD Open. 2022;8(2):e002641. McFall SL, Conolly A, Burton J. Collecting biomarkers using trained interviewers. Lessons learned from a pilot study. Surv Res Methods. 2014;8(1):57–66. Hyland A, Ambrose BK, Conway KP, Borek N, Lambert E, Carusi C, et al. Design and methods of the Population Assessment of Tobacco and Health (PATH) Study. Tob Control. 2017;26(4):371–8. Zipf G, Chiappa M, Porter KS, Ostchega Y, Lewis BG, Dostal J. Health and Nutrition Examination Survey Plan and Operations, 1999–2010. National Center for Health Statistics; 2013. Groves RM. Three Eras of Survey Research. Public Opin Q. 2011;75(5):861–71. Pashazadeh F, Cernat A, Sakshaug JW. The Effects of Biological Data Collection in Longitudinal Surveys on Subsequent Wave Cooperation. In: Advances in Longitudinal Survey Methodology [Internet]. John Wiley & Sons, Ltd; 2021 [cited 2024 Apr 9]. pp. 100–21. Available from: https://onlinelibrary.wiley.com/doi/abs/ 10.1002/9781119376965.ch5 Boyle P. others. Willingness to participate in health surveys involving biomarker data collection. J Aging Res. 2021;45(2):201–18. Sakshaug JW, Couper MP, Ofstedal MB. Characteristics of physical measurement consent in a population-based survey of older adults. Med Care. 2010;48(1):64–71. Cernat A, Sakshaug JW, Chandola T, Nazroo J, Shlomo N. Nurse Effects on Non-response in Survey-Based Biomeasures. Int J Soc Res Methodol 2021 July 4;24(4):487–99. Gavrilova N, Lindau ST. Salivary Sex Hormone Measurement in a National, Population-Based Study of Older Adults. J Gerontol Ser B. 2009;64B(suppl1):i94–105. Kumari M, Andrayas A, Al Baghal T, Burton J, Crossley TF, Jones KS, et al. A randomised study of nurse collected venous blood and self-collected dried blood spots for the assessment of cardiovascular risk factors in the Understanding Society Innovation Panel. Sci Rep. 2023;13(1):13008. Benzeval M, Andrayas A, Mazza J. others. Does the feedback of blood results in observational studies influence response and consent? A randomised study of the Understanding Society Innovation Panel. BMC Med Res Methodol. 2023;23:134. West BT, Blom AG. Explaining Interviewer Effects: A Research Synthesis. J Surv Stat Methodol. 2017;5(2):175–211. Cernat A, Sakshaug JW, Chandola T, Nazroo J, Shlomo N. Nurse Effects on Non-response in Survey-Based Biomeasures. Int J Soc Res Methodol. 2020;0(0):1–14. Clarke P, Houle J. Selection bias in biomarker data: Using Heckman models to adjust for nonresponse in HIV prevalence estimates. Demography. 2014;51(2):975–96. Hannemann T, Shlomo N, Chandola T, Chatzi G. Compensating for non-response in biosocial research: Simulation study from a cross-sectional analysis. 2018. Bossuyt PMM. Clinical validity: defining biomarker performance. Scand J Clin Lab Investig Suppl. 2010;242:46–52. Máčová L, Bičíková M, Vitamin D. Current challenges between the laboratory and clinical practice. Nutrients. 2021;13(6):1758. Khan AI, Vasquez Y, Wians FH Jr, Kroll MH. The variability of results between point-of-care testing glucose meters and the central laboratory analyzer. Arch Pathol Lab Med. 2006;130(10):1527–32. Mohamed R, Campbell JL, Cooper-White J, Dimeski G, Punyadeera C. The impact of saliva collection and processing methods on CRP, IgE, and Myoglobin immunoassays. Clin Transl Med. 2012;1:1–8. Martin AR, Kanai M, Kamatani Y, Okada Y, Neale BM, Daly MJ. Clinical use of current polygenic risk scores may exacerbate health disparities. Nat Genet. 2019;51(4):584–91. Waldmann S, Sakshaug JW, Cernat A. Interviewer Effects on the Measurement of Physical Performance in a Cross-National Biosocial Survey. J Surv Stat Methodol. 2023;smad031. Cernat A, Sakshaug JW. The Impact of Nurse Continuity on Biosocial Survey Participation. Surv Methods Insights Field SMIF [Internet]. 2020 July 7 [cited 2020 July 7]; Available from: https://surveyinsights.org/?p=13222 Cernat A, Sakshaug JW. Interviewer Effects in Biosocial Survey Measurements. Field Methods. 2021. Crimmins EM, Zhang YS, Kim JK, Frochen S, Kang H, Shim H, et al. Dried blood spots: Effects of less than optimal collection, shipping time, heat, and humidity. Am J Hum Biol. 2020;32(5):e23390. Murphy JM, Browne RW, Hill L, Bolelli GF, Abagnato C, Berrino F, et al. Effects of transportation and delay in processing on the stability of nutritional and metabolic biomarkers. Nutr Cancer. 2000;37(2):155–60. Wu T, Rifai N, Roberts LJ, Willett WC, Rimm EB. Stability of measurements of biomarkers of oxidative stress in blood over 36 hours. Cancer Epidemiol Biomarkers Prev. 2004;13(8):1399–402. Dakappagari N, Zhang H, Stephen L, Amaravadi L, Khan MU. Recommendations for clinical biomarker specimen preservation and stability assessments. Bioanalysis. 2017;9(8):643–53. McDade TW, Williams S, Snodgrass JJ. What a drop can do: dried blood spots as a minimally invasive method for integrating biomarkers into population-based research. Demography. 2007;44(4):899–925. Lindrose AR, McLester-Davis LWY, Tristano RI, Kataria L, Gadalla SM, Eisenberg DTA, et al. Method comparison studies of telomere length measurement using qPCR approaches: A critical appraisal of the literature. PLoS ONE. 2021;16(1):e0245582. Wu AHB. Biological and analytical variation of clinical biomarker testing: implications for biomarker-guided therapy. Curr Heart Fail Rep. 2013;10:434–40. Aylward LL, Hays SM, Koch HM, Jones KC, Levy L, Bevan R. Sources of variability in biomarker concentrations. J Toxicol Environ Health Part B. 2014;17(1):45–61. Arch BN, Blair J, McKay A, Gregory JW, Newland P, Gamble C. Measurement of HbA1c in multicentre diabetes trials – should blood samples be tested locally or sent to a central laboratory: an agreement analysis. Trials. 2016;17:8. Tworoger SS, Hankinson SE. Use of biomarkers in epidemiologic studies: minimizing the influence of measurement error in the study design and analysis. Cancer Causes Control. 2006;17:889–99. Han W, Li L. Evaluating and minimizing batch effects in metabolomics. Mass Spectrom Rev. 2022;41(3):421–42. Taioli E, Kinney P, Zeleniuch-Jacquotte A, Fulton H, Shore RE, Toniolo P, et al. Application of reliability models to studies of biomarker validation. Environ Health Perspect. 1994;102(3):306–9. Goh WWB, Wang W, Wong L. Why batch effects matter in omics data, and how to avoid them. Trends Biotechnol. 2017;35(6):498–507. Price EM, Robinson WP. Adjusting for batch effects in DNA methylation microarray data, a lesson learned. Front Genet. 2018;9:83. Leek JT, Scharpf RB, Corrada Bravo H, Simcha D, Langmead B, Johnson WE, et al. Tackling the widespread and critical impact of batch effects in high-throughput data. Nat Rev Genet. 2010;11(10):733–9. Lee JE, Kim SY, Shin SY. Effect of repeated freezing and thawing on biomarker stability in plasma and serum samples. Osong Public Health Res Perspect. 2015;6(6):357–62. Rebholz SL, Melchior JT, Welge JA, Remaley AT, Davidson WS, Woollett LA. Effects of multiple freeze/thaw cycles on measurements of potential novel biomarkers associated with adverse pregnancy outcomes. J Clin Lab Med. 2017;2(1):10–16966. Keshavan A, Heslegrave A, Zetterberg H, Schott JM. Stability of blood-based biomarkers of Alzheimer’s disease over multiple freeze-thaw cycles. Alzheimers Dement Diagn Assess Dis Monit. 2018;10:448–51. Mitchell BL, Yasui Y, Li CI, Fitzpatrick AL, Lampe PD. Impact of freeze-thaw cycles and storage time on plasma samples used in mass spectrometry-based biomarker discovery projects. Cancer Inf. 2005;1:117693510500100110. Wang C, Kohli N, Henn L. A Second-Order Longitudinal Model for Binary Outcomes: Item Response Theory Versus Structural Equation Modeling. Struct Equ Model Multidiscip J. 2015;0(0):1–11. Ehrmeyer SS, Laessig RH. Has compliance with CLIA requirements really improved quality in US clinical laboratories? Clin Chim Acta. 2004;346(1):37–43. Nieto Sierra JA, Gefen D. 35 years after CLIA 1988: Key insights and policy implications among laboratory professionals. PLoS ONE. 2024;19(9):e0311251. Alexander TS. Licensure and Accreditation. In: Manual of Molecular and Clinical Laboratory Immunology. 2024. pp. 363–8. Hartley SW, Mullikin JC. QoRTs: a comprehensive toolset for quality control and data processing of RNA-Seq experiments. BMC Bioinformatics. 2015;16:1–7. Sheng Q, Vickers K, Zhao S, Wang J, Samuels DC, Koues O, et al. Multi-perspective quality control of Illumina RNA sequencing data analysis. Brief Funct Genomics. 2017;16(4):194–204. Drabovich AP, Martínez-Morillo E, Diamandis EP. Toward an integrated pipeline for protein biomarker development. Biochim Biophys Acta BBA - Proteins Proteom. 2015;1854(6):677–86. Robinson P. Task– based language learning: A review of issues. Lang Learn. 2011;61:1–36. Herting MM, Gautam P, Chen Z, Mezher A, Vetter NC. Test-retest reliability of longitudinal task-based fMRI: Implications for developmental studies. Dev Cogn Neurosci. 2018;33:17–26. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 05 Nov, 2025 Reviewers agreed at journal 29 Oct, 2025 Reviewers invited by journal 29 Oct, 2025 Editor assigned by journal 28 Oct, 2025 Editor invited by journal 03 Oct, 2025 Submission checks completed at journal 02 Oct, 2025 First submitted to journal 02 Oct, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7711723","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":542228190,"identity":"a0d704aa-cee8-4df8-b133-e31fb08a0493","order_by":0,"name":"Alexandru Cernat","email":"data:image/png;base64,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","orcid":"","institution":"University of Manchester","correspondingAuthor":true,"prefix":"","firstName":"Alexandru","middleName":"","lastName":"Cernat","suffix":""},{"id":542228191,"identity":"73425e96-b17c-4ce0-88d7-3b9e4ac55981","order_by":1,"name":"Brad Edwards","email":"","orcid":"","institution":"Westat (United States)","correspondingAuthor":false,"prefix":"","firstName":"Brad","middleName":"","lastName":"Edwards","suffix":""},{"id":542228192,"identity":"5ecd93a1-8d81-4178-97b4-2043fb6542e0","order_by":2,"name":"Colter Mitchell","email":"","orcid":"","institution":"University of Michigan–Ann Arbor","correspondingAuthor":false,"prefix":"","firstName":"Colter","middleName":"","lastName":"Mitchell","suffix":""},{"id":542228193,"identity":"a204dcb0-4460-4fb0-b706-534abede4675","order_by":3,"name":"Joseph W. Sakshaug","email":"","orcid":"","institution":"Institut für Arbeitsmarkt und Berufsforschung","correspondingAuthor":false,"prefix":"","firstName":"Joseph","middleName":"W.","lastName":"Sakshaug","suffix":""}],"badges":[],"createdAt":"2025-09-25 10:23:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7711723/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7711723/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":95546899,"identity":"7206e910-7858-48a0-81bb-51334b3858dd","added_by":"auto","created_at":"2025-11-10 12:41:14","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1998290,"visible":true,"origin":"","legend":"","description":"","filename":"TSEbiomarkerv1.3.docx","url":"https://assets-eu.researchsquare.com/files/rs-7711723/v1/7616bdec6f7eec3dab51f6ea.docx"},{"id":95546896,"identity":"3acea8de-0308-4023-8e6a-590e634c2575","added_by":"auto","created_at":"2025-11-10 12:41:14","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7170,"visible":true,"origin":"","legend":"","description":"","filename":"2d1a835dcea847eab5ebabc4f156749d.json","url":"https://assets-eu.researchsquare.com/files/rs-7711723/v1/94ee10ab938ac1b5ab460e1f.json"},{"id":95654982,"identity":"5be58c3e-77f0-48ab-9d9d-506384ab0e26","added_by":"auto","created_at":"2025-11-11 16:13:57","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":164020,"visible":true,"origin":"","legend":"","description":"","filename":"2d1a835dcea847eab5ebabc4f156749d1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7711723/v1/a9323c139e9ce5f3260a4d39.xml"},{"id":95546897,"identity":"fbc72cfe-b0c5-40cd-8a0f-389b01800b22","added_by":"auto","created_at":"2025-11-10 12:41:14","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":181967,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7711723/v1/f8d800c9dea15234b6f7f832.png"},{"id":95546898,"identity":"8dd9ed26-b165-4635-93cf-386ad4cf042f","added_by":"auto","created_at":"2025-11-10 12:41:14","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":47549,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7711723/v1/4162b6b09cf77de2616d357f.png"},{"id":95546900,"identity":"7bf99fc4-11b6-4a97-9b5e-38a3bb987ff4","added_by":"auto","created_at":"2025-11-10 12:41:14","extension":"xml","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":163175,"visible":true,"origin":"","legend":"","description":"","filename":"2d1a835dcea847eab5ebabc4f156749d1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7711723/v1/63e61510f3223bc39a2ebb96.xml"},{"id":95546902,"identity":"4cb1170f-ea30-47e0-bd9e-6c59cd6f6d7a","added_by":"auto","created_at":"2025-11-10 12:41:14","extension":"html","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":172106,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7711723/v1/2e32f9acde9287828a7759a1.html"},{"id":95546895,"identity":"fee29534-ae80-4356-8a7a-536aa2f1d299","added_by":"auto","created_at":"2025-11-10 12:41:14","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":182299,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eAdapted version of the TSE framework for biomeasure data collection. Highlighted in red are new stages and errors due to the collection of biomeasures\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7711723/v1/e5379333002cd53d5ea5fd11.png"},{"id":95659909,"identity":"4443928b-b24c-44bb-8841-e620c9a8cb87","added_by":"auto","created_at":"2025-11-11 16:29:51","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1232799,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7711723/v1/e7e461ca-ba03-4aba-bed7-ad130361a14b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Total Survey Error in Biomeasure Data Collection","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eIn recent years, the integration of biomeasure data collection into social surveys (so-called \u0026ldquo;biosocial surveys\u0026rdquo;) has significantly expanded, leading to important new research that leverages the strengths of probability sampling from the general population and the precision of biological measurements. Nevertheless, combining social surveys with these measurements brings unique challenges regarding data quality. The Total Survey Error (TSE) framework has been traditionally used by survey researchers to guide decision-making processes regarding data collection and the evaluation of data quality (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). However, the increased complexity and the new actors involved in collecting bio-social survey data necessitate adapting this framework.\u003c/p\u003e\u003cp\u003eBiomeasure collection, encompassing measures from anthropometrics to molecular biomarkers, presents new sources of potential error not typically encountered in traditional social surveys. These errors arise from the complexity of collecting biomeasures outside of controlled clinical environments, involving multiple actors such as nurses and lay interviewers. Each stage of this process, from initial sampling to post-survey adjustments, introduces opportunities for error, thus complicating the trade-offs necessary to maximise data quality. Due to the complex nature of these errors, adapting the TSE framework to encompass biomeasure collection is crucial. Traditional TSE components like noncoverage, nonresponse, measurement, and processing errors must be re-evaluated and expanded to include the challenges of collecting biomeasures. For instance, the potential for nonresponse (or nonconsent) bias increases when participants are reluctant to provide invasive biological samples or are too burdened to participate in any physical data collection. Further, implementing biomeasure collection outside clinical settings is inherently complex and error-prone. Logistics such as equipment provision, training, specimen collection, transportation, and processing require rigorous protocols to ensure high-quality data collection. These steps, including the involvement of various data collection personnel with differing levels of expertise and in different environments, can introduce variability, impacting the reliability and validity of the collected data.\u003c/p\u003e\u003cp\u003eThis article expands on the TSE framework to include the complexities of collecting biomeasures in social surveys, highlighting their unique challenges. By examining the existing components of the TSE framework as well as proposing new stages and types of errors, we provide a guide for researchers collecting biomeasures as part of social survey data collection.\u003c/p\u003e"},{"header":"2. Overview","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Typology of Biomeasures\u003c/h2\u003e\u003cp\u003eThe integration of biomeasures into social and health research, from anthropomorphic measures to glucose to brain structure to molecular measures, has allowed researchers to examine physical and biological mechanisms, precursors to disease, residues of exposure, and moderators of response (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). Scholars who study human health and development have been trying to integrate biological theory, methods, and data for centuries (\u003cspan additionalcitationids=\"CR5 CR6 CR7\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). And since the late 20th century, there has been a rapid integration of biological measures into social and health research, in part due to the rapid technological advancements that afforded these measurements (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). Several rationales have been provided for integrating bio-measures with social and behavioural data (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eA broad definition of a biomeasure is an objectively measured indicator of an organism, its functioning, or its systems (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Because biomeasures have such a wide range of potential uses, they can often be challenging to classify. For example, for some researchers, biology is expected to be a mechanism linking exposures to health (i.e., the way for the environment to \u0026ldquo;get under the skin\u0026rdquo;), but they are also of interest as indicators of physical health status, exposure and disease, so that, even if they are not part of the causal pathway from external stimuli to outcome, researchers can reconstruct the past and potentially predict the future (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). Indeed, each system could have potentially hundreds of biomeasures, some of which are more typically used as causal mechanisms and others as predictive indicators. Similarly, biomeasures could be classified by how they are collected, in what tissue they reside (if any), or what function they measure.\u003c/p\u003e\u003cp\u003eMost biomeasures were not developed with the goal of examining variation related to social behaviour, to study mechanisms linking the environment to child development outcomes, or even to be measured outside of a lab or clinic. Furthermore, most were developed on homogeneous clinical populations that are typically higher income, western, white, and European (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). As such, their use in social surveys needs to account for new types of errors that might arise. Although far from exhaustive, by looking at a number of large population-based studies (e.g., National Study of Adolescent to Adult Health; Future of Families and Child Well Being; Health and Retirement Study), biomeasure networks (e.g., Early Genetics and Lifecourse Epidemiology \u0026ndash; EaGLE; Environmental Influences on Child Health Outcomes; NIA Biomarker Network), and biomeasure reviews (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e), we discuss key biomeasures likely to be included in survey-based research studies (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eMain types of biomeasures and their data collection characteristics\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBiomeasure\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCollection mode\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCollection cost\u003c/p\u003e\u003cp\u003e1 Low \u0026minus;\u0026thinsp;5 High\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eSusceptibility to shipping effects\u003c/p\u003e\u003cp\u003e1 Low \u0026minus;\u0026thinsp;5 High\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eCommon health measures analyzed\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eComment\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eKey Survey Methods Citations\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSaliva\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eInterviewer or self\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eDNA, RNA, microbiome, hormones, some drug use\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eWell documented costs and issues\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eAdam \u0026amp; Kumari, 2009; Zhang, Xiao, \u0026amp; Wong, 2009; Gavrilova \u0026amp; Lindau, 2009; Kaczor-Urbanowicz et al., 2017\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBlood Venous\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePhlebotomist or nurse\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eLargest potential number of tests: DNA, RNA, hormones, proteomics, metabolomics, lipids, CBC, blood chemistry, inflammation, organ function, cholesterol, glucose\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eExpensive to collect and ship, timing and temperature sensitive,\u003c/p\u003e\u003cp\u003emost versatile post collection\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eGiavarina \u0026amp; Lippi, 2017; Crimmins, Vasunilashorn, Kim, \u0026amp; Alley, 2008; Dagur \u0026amp; McCoy, 2015\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDried blood spots\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eInterviewer or self\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eReduced number, but similar types as venous\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eDifficult to self-collect\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eMcDade, Williams, \u0026amp; Snodgrass, 2007; Ostler, Porter, \u0026amp; Buxton, 2014; Jacobson et al., 2023; El-Sabawi et al., 2024; Cella et al., 2010\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHair\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eInterviewer or self\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eHormones, toxicant exposure, drug use\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eNot all respondents are eligible if hair is treated or too short\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eLawes, Hetschko, Sakshaug, \u0026amp; Eid, 2024; Ford, Boch, \u0026amp; McCarthy, 2016; Pragst \u0026amp; Balikova, 2006\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePhysiological functioning and anthropometric\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eInterviewer\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2\u0026ndash;4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eGait speed, grip strength, pulse rate, temperature, blood pressure, lung capacity, sit-stand test, vision test, hearing test, smell test\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eEquipment can be costly or inexpensive, depending on the specific measure\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eUlijaszek \u0026amp; Kerr, 1999; Casadei \u0026amp; Kiel, 2019; Gross, Jones, \u0026amp; Inouye, 2015\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eStool\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSelf\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eMicrobiome\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eCollections range from simple wipe collections to larger samples that allow for greater flexibility in use later.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eLoftfield et al., 2016; Sinha et al., 2016; Li et al., 2023\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUrine\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSelf\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eDrugs, toxicant exposure, hormones\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eFairly easy to collect, but many labs are not interested in holding\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eJohn et al., 2016; Fenton, Johnson, McManus, \u0026amp; Erens, 2001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTask-based cognition\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eInterviewer or self (via app)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eCognition, memory\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eCan be difficult to administer\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eChalhoub-Deville, 2013; Hollnagel, 2003; Suzuki et al., 2004\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eEach biomeasure has strengths and weaknesses, some of which are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Sources of measurement error are specific to the biomeasure collection method and play a significant role in the quality of the final estimates. In the table, we provide eight common biomeasures and discuss four broad categories of issues: 1) who collects the biomeasure, 2) relative collection cost, 3) potential issues, and 4) common health measures analyzed (though far from an exhaustive list).\u003c/p\u003e\u003cp\u003eVenous blood collection provides a comprehensive sample that is suitable for the widest range of analyses. However, collecting venous blood in a home setting often requires a healthcare professional due to the invasiveness of the procedure, which can increase costs and reduce the convenience factor. And often, the initial processing in the field and cold storage shipping can be expensive if it is far from the processing lab. Dried blood spots (DBS) offer a practical alternative for home collection of blood samples, as they are stable at room temperature and can be easily mailed. The downside is that the volume of blood obtained is limited, which may affect the breadth of assays that can be performed, and extraction from DBS can sometimes be less efficient than from whole blood. And there are some known biases resulting from temperature and humidity exposures. Also, although DBS could be collected by the individual, evidence suggests this is challenging for most participants\u0026ndash;with often low-quality samples. Newer devices that try to bridge the gap between venous blood and DBS, such as the TASSO device, are being tested more regularly now.\u003c/p\u003e\u003cp\u003eAnother minimally invasive mode of collection is saliva collection. It does not require specialized equipment and is well-suited for hormone (e.g., cortisol) and genomic studies in particular. Because it is simple to collect and non-invasive, it is ideal for use in large-scale self-collection efforts and in pediatric populations. However, the reliability of saliva samples can be influenced by recent eating or drinking and contamination with food particles or blood. And some hormones require time-specificity in the collection.\u003c/p\u003e\u003cp\u003eAnthropometric measures, such as weight, height, and waist circumference, can be self-collected and serve as indicators of nutritional status or disease risk, but the accuracy of self-reported measurements can be compromised by a lack of precision or user error. Many of the other physiological assessments must be performed by a trained interviewer. Materials needed for these assessments can range from relatively inexpensive (a scale, tape measure, or stopwatch for each interviewer) to far more expensive devices, depending on the biomarker of interest.\u003c/p\u003e\u003cp\u003eHair samples provide a long-term biomarker analysis option, reflecting exposure to substances like heavy metals or drugs over months (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). They are relatively easy to collect (by an interviewer), ship, and store without the need for special conditions. However, external contamination (from treatments, dyes) can influence some measures, and participants with very short hair or semi-permanent procedures, or sensitivities around their hair, often must be excluded.\u003c/p\u003e\u003cp\u003eCollecting stool samples at home is advantageous for microbiome studies, detecting gastrointestinal conditions, and parasite testing. The process is non-invasive, and samples can be mailed to the laboratory for analysis. However, the principal challenge is ensuring proper sample preservation and avoiding degradation, which can affect bacterial viability and microbial community analysis. Moreover, the collection process might be uncomfortable or unappealing for some participants, potentially impacting participation rates. Urine collection, on the other hand, is more straightforward, non-invasive, and ideal for detecting metabolites, hormones, and drug substances. Yet, variability in concentration due to hydration levels and the potential for sample contamination are recognized limitations. Also, the shipping can be particularly challenging if an interviewer isn\u0026rsquo;t taking it with them. Moreover, as with stool samples, fewer labs are equipped to work with these samples.\u003c/p\u003e\u003cp\u003eTask-based cognition tests have been developed in clinical settings to assess a wide range of brain functions (memory, attention, and processing speeds, executive function, language and verbal fluency, and visuospatial ability) in various fields, including psychology, neurology, and gerontology (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). Many have been applied in surveys, especially in child studies and surveys of elderly populations. Structured tasks or activities are typical administration forms. Some task-based cognition tests can be self-administered, but others require interviewer administration to ensure standard conditions are met (such as adequate lighting and distance for a vision test) or to deploy specialized, costly devices (such as equipment for a smell test).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2 Models for Biomeasure Data Collection in Survey Research\u003c/h2\u003e\u003cp\u003eA range of models have been implemented for combining biomeasures and survey data collections (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). In the study design phase, new studies often have a wider range of potential biomeasure models to consider than studies in a later phase seeking to add biomeasure components to existing surveys (such as a long-running panel study). Tradeoff dimensions include administration mode, data collection personnel, collection location, processing location and procedures, and timing of consent and collection. Some studies may be created with biomeasure collection as a primary focus, and with survey collection as secondary or supplemental. The National Health and Examination Survey (NHANES) in the U.S. is one example. Much more common are established surveys (repeated cross-sectional or longitudinal) that add biomeasure components midstream (such as the UK Household Longitudinal Study or the English Longitudinal Study of Ageing). In either case, tradeoffs at each stage have cost and quality considerations, and extending the TSE framework can be helpful in thinking about biomarker collection and processing.\u003c/p\u003e\u003cp\u003eBeebe (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e) suggests three models for biomeasure collection in surveys: self-administered, interviewer-administered in the home, and administered by an interviewer or medically trained person in a clinic. Simpler biomeasure data (e.g., hair, saliva, urine, accelerometry) can be collected by participants themselves (\u003cspan additionalcitationids=\"CR21\" citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e). Other biomeasures (e.g., blood, grip strength) require collection by trained staff in-person, often with special equipment or tools. Height and weight are interesting examples because self-reports are sometimes deemed too unreliable to include in a data set on their own, and are obtained by direct measurement in person. Skill requirements differ \u0026ndash; field interviewers can be trained to collect some physical and cognitive functioning data (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e), but others (e.g., venous blood) require a phlebotomist or nurse (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). If tests require highly skilled personnel (e.g, physicians, radiologists), the home may not be an appropriate setting, and travel to a clinic or other well-equipped location may be the only choice. Some studies have considered mobile vans as a compromise, with NHANES being a notable example (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). Bio-specimens may need to be shipped to a laboratory for processing under controlled conditions with specialized equipment and kept at controlled temperatures.\u003c/p\u003e\u003cp\u003eTiming in relation to the data collection has two dimensions: time from the interview, and repetition across interviews on longitudinal or repeated cross-sectional surveys. Some biomeasure data is so tightly linked to the survey that it must be collected at the same time or very close to it. For example, the Population Assessment of Tobacco and Health (PATH) Study collects data about tobacco and drug use in an interview, and schedules a blood draw within 10 days to compare values from the blood with the self-reports, and to augment survey reports on health behaviors (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). Other data that are more stable over time may be collected in a looser relationship to the interview, for example, altimetry data measuring activity levels for a week. The other dimension is about repeating measures over time. For longitudinal and repeated cross-sectional surveys, it is often critical to collect the same measure repeatedly under the same conditions. Controlling the administration and processing of biomeasure data takes on even more importance in such comparative work.\u003c/p\u003e\u003cp\u003eCombining biomeasure collection with survey collection can add additional error sources at every point in the process (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e) due to use of nurse or phlebotomist, collection equipment, shipment, and laboratory. Each activity in the flow from design, training, and collection through post-collection processing adds friction that can lead to error.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Adapted TSE for Biomeasure Data Collection\u003c/h2\u003e\u003cp\u003eTo better understand the implications of collecting biomeasure data in surveys, we propose to expand the TSE framework. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e builds on the traditional TSE framework while highlighting the new stages as well as their related errors.\u003c/p\u003e\u003cp\u003eOn the selection side (right-hand side of Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), in addition to the typical survey selection stages (such as target population, sampling frame, and sample), there are further stages for biomeasure respondents. This group can be significantly different from regular survey respondents due to eligibility criteria, consent requirements, and respondents\u0026rsquo; abilities and motivation. Moreover, some biomeasures may need to be shipped to a lab or research center for processing or analysis. This stage can also lead to loss of data and selection, for example, due to items packed in dry ice but delivered too late to maintain the proper temperature.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eOn the measurement side (left-hand side of Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), once the biomeasures are collected, the handling and shipment of them can impact their integrity and the quality of the measurements. This is especially the case for biological samples (e.g., tissue, blood), which may need to be sent to specialized labs. If the shipments take too long or if the packaging is not adequate, it can have an impact on the measurements. Additionally, once the specimens arrive, the lab and the batches in which they are analysed can influence subsequent results. In the next sections, we discuss each stage of the revised TSE diagram and highlight how these stages and errors can affect the collection and quality of biomeasure collections.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. TSE Stages in Biomeasure Research: Selection","content":"\u003cp\u003eThis section discusses biomeasure error in the selection process, from the target population to the sampling frame, from the frame to sample selection, from sample selection to respondents, from respondents to raw data, and from raw data to post-survey adjustment. For each of these stages, we discuss issues that are specific to surveys with biomeasure data collection.\u003c/p\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e3.1 Target Population to Sampling Frame: Coverage Error\u003c/h2\u003e\u003cp\u003eThe first stage of selection identifies the population of interest and the sampling frame that will be used. The difference between these two leads to coverage error. The size of the error depends on the proportion of people missing from the sampling frame and how different they are from the general population in terms of the variables of interest.\u003c/p\u003e\u003cp\u003eIf the survey design starts from a probability sample of the general population, followed by biomeasure data collection, then coverage error is similar to that present in social surveys. Additional errors could appear if there is a relationship between the likelihood of people being in the sampling frame and the biomeasures of interest. For example, people with poorer health, and thus different levels of biomeasures, might be less likely to have a phone (and not be covered by a random-digit-dial sampling procedure) or registered to vote (if voting records are used for sampling). If the data collection is based on a non-probability selection process, such as the UK Biobank or the US All of Us Research Program, some people will have no chance of being selected. In this case, it is difficult to separate coverage error from non-response error (see next section), leading to selection bias. This can be especially problematic if the process underrepresents important groups such as ethnic minorities.\u003c/p\u003e\u003cp\u003eResearchers also need to consider whether the definition of the target population needs to be adjusted due to the collection of biomeasures. This is especially the case if some biomeasure sub-groups have no chance of being selected. For example, cortisol from hair samples cannot be collected from men who have very short hair or are bald. In this case, either the target population is defined in such a way as to exclude such people or alternative measures need to be considered (e.g., using saliva or urine).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e3.2 Sampling Frame to Sample: Sampling Error\u003c/h2\u003e\u003cp\u003eThe next stage in data collection is using a sampling frame (a list of the population of interest) or a procedure to select households or individuals. If the selection process is random, it can lead to a degree of uncertainty, which is known as sampling error. The size of the sample and the procedure used can impact the amount of sampling error. When collecting biomeasure data as part of a survey, the size of the sampling error is not different unless the procedure is influenced by the biomeasure collection. This could happen if, for example, biomeasure information is collected under a stratified or clustered sample design.\u003c/p\u003e\u003cp\u003eIn studies where cases are selected using a non-probability selection process, there is often a tradeoff between sampling error and systematic bias due to coverage and non-response. For example, a large study such as the UK Biobank will have apparently precise estimates due to the large sample size. Nevertheless, the estimates may be biased, with systematic selection biases due to coverage and non-response errors. One of the reasons for the increasing popularity of combining probability samples with biomeasure data collection is to avoid the perils of such biased samples.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e3.3 Sample to Survey Respondents: Survey Nonresponse\u003c/h2\u003e\u003cp\u003eThe next stage in the selection process is collecting data from all sampled cases. When some individuals decide not to participate in the study, there can be a nonresponse error. As in the case of coverage error, the amount of error depends on the proportion of cases that do not participate and how different they are from the participants. There are numerous mechanisms that can affect the likelihood of participating in a survey that involves biomeasure data collection. While some individuals may be averse to participating in such a study, others may find the study\u0026rsquo;s focus to be interesting and salient to their lives, which may encourage them to participate. Research has shown that the study\u0026rsquo;s topic plays an important role in influencing survey participation, in addition to other survey design features, such as the offering of incentives and data collection mode (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eBiomeasure data collection can also have detrimental effects on retention in longitudinal studies because of the extra respondent burden. In a quasi-experimental study in the UK, Pashazadeh et al. (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e) exploited a scenario where certain participants were randomly selected for a nurse visit for biological data collection. They found a short-term negative effect on subsequent wave cooperation (wave 3) for those included in the nurse visit, suggesting that the increased burden led to a slight decrease in participation rates following the visit. However, this effect was not observed in later waves (\u003cspan additionalcitationids=\"CR5 CR6\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e), indicating that the initial negative impact of the nurse visit on survey cooperation did not persist over time.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e3.4 Survey Respondents to Biomeasure Respondents: Biomeasure Nonresponse\u003c/h2\u003e\u003cp\u003eIn addition to the nonresponse mechanisms present in social surveys, collecting biomeasure data can introduce additional sources of nonresponse error. Firstly, some studies collect biomeasures at multiple stages (e.g., a survey followed by a separate biomeasure data collection interview) undertaken by different actors, e.g., nurses. This could lead to additional nonresponse. Secondly, biomarkers often need specific consent before the measure can be collected, which can be either verbal or written. Consent requests can increase the amount of nonresponse. Third, the collection of some biomeasures might not be possible, especially for certain respondents. For example, walking speed cannot be collected for non-ambulatory participants; hair cannot be collected if participants are bald or have dyed hair; it might be impossible to collect full blood from some participants; and a vision test cannot be conducted with people in advanced stages of Alzheimer\u0026rsquo;s disease. The mechanism for the nonresponse in these cases depends both on the respondent's characteristics and on the interviewer/nurse\u0026rsquo;s skill. Finally, for many studies, respondents need to agree to allow the biomeasure data or results derived from it to be archived and shared with other researchers for future analyses.\u003c/p\u003e\u003cp\u003eThere is some evidence regarding nonresponse that is specific to biomeasure data collection. For example, willingness to participate in biomeasure data collection varies considerably by topic. Boyle and colleagues (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e) reported that while about 77% of respondents showed willingness to participate in a less burdensome survey on an interesting topic, only 62%-64% expressed willingness to participate in more invasive health surveys involving measurements like blood pressure or waist circumference, and just 45% were open to a finger stick blood draw, highlighting potential nonresponse biases in surveys with biomeasure due to their intrusiveness and the burden they impose. Among those who completed the Adult Interview in an actual biomeasure collection study, the weighted response rates 63.6% for providing a urine sample and 43.0% for providing a blood sample, providing further evidence of an intrusiveness effect.\u003c/p\u003e\u003cp\u003eIn another study, Sakshaug et al. (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e) examined factors influencing consent to physical measurements in the Health and Retirement Study, focusing on older adults. The study found that while 79% of respondents consented to all physical measurements, factors such as demographics, health status, levels of survey cooperation, and interviewer characteristics significantly influenced consent rates. Key findings indicated that Hispanics paired with interviewers bilingual in Spanish and English, people with recent doctor visits, and people with diabetes were more likely to consent, whereas younger individuals, people with functional limitations, or people who were reluctant survey participants were less likely to consent to biomeasure collection. There is also some evidence regarding biomeasure nonresponse patterns in general population surveys. For example, in large panel studies in the UK, older individuals, those with higher educational attainment, homeowners, and white respondents generally showed higher participation in biomeasure collection (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). Conversely, women were less likely to provide blood samples despite their higher engagement in initial stages like nurse visits. Household dynamics, such as living alone and having a partner, can affect participation across different stages. Additionally, both long-term illness and good to excellent self-reported health positively influenced participation, underscoring how personal health status and socioeconomic factors shape engagement in such studies.\u003c/p\u003e\u003cp\u003eThe National Social Life, Health, and Aging Project\u0026rsquo;s (NSHAP) implementation of in-home salivary specimen collection from older adults reported that 90.6% of participants provided saliva samples, with a significantly lower participation rate among Black participants compared to other racial groups (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). Interestingly, factors such as age, gender, education, and self-rated health did not significantly influence the likelihood of providing a saliva sample, highlighting specific demographic patterns in the participation of older adults in biomeasure data collection within large social surveys.\u003c/p\u003e\u003cp\u003e\u003cb\u003eImpact of data collection design on participation\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThere is some evidence for differences in participation based on the mode of the biomeasure collection. An experimental study in the UK showed that self-completion of dried blood spots had significantly lower participation compared to that done by a nurse (35% vs. 74%), although there were no differences in participant characteristics (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). There is also limited research on the impact of offering feedback as a way to encourage participation in biomeasure collection. Benzeval et al. (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e) have shown that there is no effect of offering feedback on the likelihood of collecting a blood sample in a panel study in the UK, but it did have an effect on consent to giving blood. There was also some evidence of differential effects with the highest impact on web participants.\u003c/p\u003e\u003cp\u003eWhile design decisions such as mode and giving feedback aim mainly to maximise participation and minimise costs, they may also have unintended consequences on the measurement and behaviour. For example, receiving feedback about their health could encourage people to exercise or change their eating habits. From a research perspective, this is a particular concern in longitudinal studies and could lead to a kind of conditioning effect unique to biomeasure collection.\u003c/p\u003e\u003cp\u003e\u003cb\u003eInterviewer/nurse impact on non-response\u003c/b\u003e\u003c/p\u003e\u003cp\u003eNonresponse can also be influenced by the interviewer's skills and characteristics, as highlighted by West and Bloom (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). This can also be the case for biomeasure collection. Looking at two waves of the UK Household Longitudinal Study and ELSA, researchers found that nurses accounted for between 5% and 14% of the variance in nonresponse rates to biomeasure collection. The impact of the nurses varied by stage of data collection, with 5%, 9%, and 14% of the variation explained for nurse visits, consent to blood, and blood collection. There was some evidence of effects of nurse characteristics, with older nurses generally being more successful at obtaining cooperation and consent for data collection, while nurses with more survey experience were more likely to successfully collect blood samples (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eFor longitudinal studies with multiple waves of biomeasure data collection, previous research shows that switching nurses between survey waves does not significantly impact respondents' participation in biomeasure collection. This includes participation in the nurse visit, consent to blood collection, and successful blood collection (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e). Furthermore, in some cases, it can be beneficial, especially for respondents who did not participate in the previous wave's biomeasure collection. This suggests that nurse continuity is not critical for ensuring high participation rates in subsequent waves of biomeasure collection.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e3.5 Biomeasure Respondents to Raw Data: Data Loss Errors\u003c/h2\u003e\u003cp\u003eThe process of converting a biological specimen to usable data is complex and has several steps: transfer from the field to the lab, inventorying an intermediate or analytic lab, processing, assay analyses, and quality control assessments. Each step provides an opportunity to lose, contaminate, or modify samples. However, most research on lab quality focuses on issues of labs influencing measurement error rather than representation. There are at least two major areas of concern for representation. First, loss of sample, which reduces statistical power, and if that loss is correlated with a variable of interest, bias can result. Because of the cost of obtaining the sample in the first place, the price of losing the sample is high. For that reason, most studies employ detailed shipping and storage protocols. And many labs will rerun samples that fail quality control so as to obtain as much information as possible. Relatedly, the number of times a sample is transferred to another lab or location, the more likely it is to be lost or unintentionally modified. A second, more likely scenario, is that initial protocols are modified to fix discovered flaws, and if those early participants are more likely to be from certain populations (i.e., more compliant, the first geographic target, etc), then there could be a bias in who will provide a viable sample.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e3.6 Postsurvey Adjustment: Adjustment error\u003c/h2\u003e\u003cp\u003eAs the previous sections have highlighted, the mechanisms of nonresponse in biomeasure data collection can be different from those of participating in a social survey and answering questions. As such, the strategies for correcting for nonresponse in biomeasure data should account for that. One strategy is to separate each stage of nonresponse and create conditional models that explain the causes of missingness for each stage. For instance, Cernat et al. (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e) developed regression models to analyze participation in nurse visits, consent to collect blood data, and the actual blood data collection. Their analysis revealed distinct predictors for each mechanism. When generating multiple weights to adjust for each mechanism, they observed no significant differences in the impact on point estimates for pulse and C-reactive protein (CRP). However, incorporating adjustments for biomeasure data collection yielded estimations that differed from those derived solely from using survey nonresponse weights. Many large studies like the UK Household Longitudinal Study, English Longitudinal Study of Ageing, Health and Retirement Study, or the China Health and Retirement Longitudinal Study provide nonresponse weights that account for some of the selection processes in biomeasure data collection.\u003c/p\u003e\u003cp\u003eOther models can also be used to correct for nonresponse in biological data collection. For example, Clarke and Houle (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e) employed Heckman selection models to adjust for nonresponse biases in biomeasure data from HIV prevalence studies. This method was specifically used to correct for selection biases due to differential nonresponse in demographic and health surveys. The Heckman model provided a way to estimate the correlation between the likelihood of participating in HIV testing and the actual HIV status, thereby correcting the HIV prevalence estimates to better reflect the true population values.\u003c/p\u003e\u003cp\u003eDifferent strategies for adjusting for nonresponse in biomeasures were explored in a simulation study using data from the English Longitudinal Study of Ageing (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e). In it, researchers examined the effects of missing data on socio-economic analyses using the C-Reactive Protein biomarker. The study explored three missing data mechanisms\u0026mdash;Missing Completely at Random (MCAR), Missing at Random (MAR), and Missing Not at Random (MNAR)\u0026mdash;and tested four compensation methods: inverse propensity weighting, a Heckman selection model, multiple imputation, and a hybrid approach combining the Heckman model with multiple imputation. The findings indicated that multiple imputation is the most effective method for addressing missing data across all scenarios, making it a robust choice for biosocial research.\u003c/p\u003e\u003cp\u003eBest practices regarding correction for nonresponse vary by discipline and journals. Some studies have no weights available to deal with nonresponse in biomeasure data. In such cases, users should consider creating their own or using alternative methods, such as multiple imputation. Even when weights do exist, some researchers may decide not to use them due to the extra complexity involved or uncertainties regarding best practices.\u003c/p\u003e\u003c/div\u003e"},{"header":"4. TSE Stages in Biomarker Research: Measurement","content":"\u003cp\u003eWe next discuss the second branch of the Total Survey Error framework: measurement. Here, we review how the process of going from the construct of interest to the raw data can be different when collecting biomeasures. We also introduce new types of errors: transmission and lab and batch effects, that can be especially important when collecting biological data.\u003c/p\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e4.1 Construct to Measurement: Validity\u003c/h2\u003e\u003cp\u003eThe extent to which a biomeasurement is valid, or captures what the researchers intend, is highly dependent on the concept of interest. Some biomeasures have been extensively tested, while others are more novel and have had limited validation. Typically, the focus is either on validating using a gold standard clinical measure (e.g, interview measured weight vs. a physician measured weight) or using the predictive capacity of a health or behavioural outcome (clinical or predictive validity (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e)). When a clinical comparison is unavailable, criterion validity is commonly used by examining whether the biomeasure correlates with known covariates. Typically, significant effort is made during the assay or protocol development to examine the validity of the biomeasure, but with more novel measures, or those biomarkers with a large number of potential algorithms (e.g., genetics), few validity analyses have been conducted.\u003c/p\u003e\u003cp\u003eOne tension for survey-paired biomarkers is that biomeasures are typically compared to a parallel test conducted in an ideal context (i.e., a clinic) vs. in-home, yet often the purpose of many studies is to provide more naturalistic measures. So, a critical question is whether an observed deviation from the gold standard is evidence of lower validity or potentially a more real-life measurement. The degree to which contextual effects lead to a systematic change in the measurement can lead to a lack of validity.\u003c/p\u003e\u003cp\u003eLack of validity of biomeasures can substantially undermine their utility in research contexts, leading to misleading scientific conclusions. One major issue is the over-reliance on biomeasures without adequate validation. Many biomeasures are reported based on initial discovery studies but lack subsequent rigorous validation in independent, representative cohorts. Such premature reliance can result in false assumptions about their predictive or diagnostic power. Another prevalent misuse concerns the lack of consideration for confounding variables and proper standardization across studies. Biomeasurements such as C-reactive protein (CRP) and interleukins can be influenced by a variety of non-disease factors, including age, sex, diet, and circadian rhythms. Without adjusting for these confounding factors, studies may draw erroneous associations between biomeasures and disease states.\u003c/p\u003e\u003cp\u003eThe lack of validity often stems from commercial pressures and the premature commercialization of biomarker assays. While the commercialization of biomeasure tests has brought innovative diagnostic tools to research, it has also led to the widespread availability of tests that may not have undergone thorough validation. For instance, genomic biomarkers such as polygenic scores, epigenetic clocks, and telomere length are often put on the market based on limited evidence, which can result in their misuse.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e4.2 Measurement to Response: Measurement Error\u003c/h2\u003e\u003cp\u003eOnce the measurement instruments are developed to capture a concept of interest, they have to be applied to collect a response from study participants. If the instrument does not capture the intended measure, this results in a measurement error. In the case of biomeasures, measurement error can occur in several ways. Instruments might have been developed and pre-tested in ideal contexts and environmental factors, such as the biomeasure collection location and timing, that can lead to some degree of measurement error (\u003cspan additionalcitationids=\"CR40\" citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). Measurement error could also arise from fine-tuning instruments and clinical cut-offs that were developed in ideal conditions but do not meet in the real world, or that were based on specific populations (e.g., white males) that may not be appropriate for generalizing to more heterogeneous populations (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cb\u003eInterviewer/nurse effects on biomarkers\u003c/b\u003e\u003c/p\u003e\u003cp\u003eAnother cause of measurement errors could be due to the interviewers. For example, a large study looking at eight waves of Survey of Health, Ageing and Retirement in Europe (SHARE) revealed moderate-to-large interviewer effects on physical performance measures in a cross-national survey, with variance partition coefficients (VPCs) ranging from 0.05 to 0.28 for biomeasures like timed chair stand, walking speed, grip strength, and peak flow, in stark contrast to the minimal effects observed for self-reported measures like height and weight (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). Notably, peak flow demonstrated the highest susceptibility to interviewer effects, followed by chair stand, walking speed, and grip strength. Additionally, significant cross-national differences in interviewer effects were identified, with particular biomeasures showing unexpectedly high VPCs in certain countries, aligning with cross-national variations found in other survey measures. The analysis also highlighted the absence of a consistent trend in interviewer effects across data collection waves, suggesting that increased interviewer experience or training over time does not uniformly improve data quality for all biomeasures, and in some cases, might even introduce variability in measurement errors.\u003c/p\u003e\u003cp\u003eA similar study based on data from two UK surveys, the English Longitudinal Survey of Ageing and the UK Household Longitudinal Study, assessed how nurses influence the measurement of fourteen different health measures, including weight, height, pulse, grip strength, and lung capacity (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). They found a medium overall effect of nurses on measurement, with nurses accounting for around 13% of the variance in all measures combined. The effect varies significantly among different measures, ranging from approximately 2% to 25%. Specifically, grip strength and lung capacity measurements are more substantially influenced by nurses compared to other measures like height, weight, and pulse. Furthermore, the study reveals that characteristics of nurses only explain a very small proportion of the observed variance, suggesting that more research is needed to understand the mechanisms behind nurse-related measurement errors.\u003c/p\u003e\u003cp\u003eA different study focusing on interviewer biological data collection in the US study National Social Life, Health, and Aging Project (NSHAP) showed large variation of the interviewers on biological measures (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). Touch sensory tests exhibit especially high interviewer effects, with about 30% of variation attributable to interviewers, while smell tests and timed balance/walk/chair stands show moderate interviewer effects around 10%. The presence of third parties or the type of dwelling did not explain this observed variation.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003e4.3 Response to Specimen: Transmission Error\u003c/h2\u003e\u003cp\u003eShipping conditions can also introduce significant measurement errors in the assessment of common biological samples by altering their integrity. Variations in temperature and delays during the transportation of blood and serum samples can lead to the degradation of sensitive biomarkers such as cytokines and hormones. The lack of precise temperature control can cause hemolysis, proteolysis, and other biochemical changes that skew biomarker concentrations. These changes may result in false-negative or false-positive results, complicating clinical diagnostics and patient management.\u003c/p\u003e\u003cp\u003eThe handling and packaging protocols during shipping also play a crucial role in reducing measurement errors. Improper handling methods, such as inadequate insulation or poor packaging materials, can exacerbate the degradation of samples. Steps like using dry ice for freezing samples, employing temperature loggers to monitor conditions, and ensuring rapid transit times are critical for preserving the integrity of biomarkers such as lipids and proteins. Consistently following standardized shipping protocols can thus minimize pre-analytical variability and ensure more accurate and reliable biomarker measurements.\u003c/p\u003e\u003cp\u003eMoreover, the duration of storage and transit has a direct impact on the stability of biomeasurements. Extended shipping times, especially under suboptimal cooling, can lead to the denaturation or degradation of biomarkers like insulin and vitamins, affecting their measurable concentrations (\u003cspan additionalcitationids=\"CR47 CR48\" citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e). These studies underscore the importance of minimizing the time between sample collection and processing to mitigate these effects. Utilizing expedited shipping services and well-maintained cold chain logistics can significantly reduce measurement errors associated with shipping, thereby enhancing the reliability of biomeasure data used in clinical and research settings.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003e4.4 Specimen to Raw Data: Lab and Batch Effects\u003c/h2\u003e\u003cp\u003eLaboratory conditions and procedural variability significantly influence the measurement error of common biomarkers such as glucose, cholesterol, inflammatory markers, molecular measures, and hormones (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e, \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e), even after accounting for pre-lab variables, including sample collection and handling. Studies highlight that different analytical techniques, reagent lots, and calibration procedures can yield varying results for the same biomarker. For instance, the accuracy and precision of enzyme-linked immunosorbent assay (ELISA) kits can differ between manufacturers, leading to inconsistent detection levels of biomarkers like C-reactive protein (CRP) and telomere length. These variations underscore the critical need for standardization of analytical methods and the implementation of rigorous quality control protocols to reduce measurement errors.\u003c/p\u003e\u003cp\u003eInter-laboratory variability also remains a significant challenge. Studies have shown that even standardized methods can produce different results when conducted across different laboratories due to differences in equipment calibration, technician expertise, and environmental conditions (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e). For example, when testing for glycosylated hemoglobin (HbA1c) and cholesterol, differences in laboratory practices can lead to significant divergence in results, potentially impacting the diagnosis (\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e). Therefore, harmonization of methods and adherence to stringent inter-laboratory proficiency testing are essential measures to minimize errors and ensure reliable biomeasurement across diverse clinical settings (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e). For many biomeasurements, harmonisation of standards and protocols has not yet occurred\u0026ndash;particularly for field-collected studies.\u003c/p\u003e\u003cp\u003eA related potential bias comes from batch effects. Batch effects refer to systematic technical variations that occur when samples are processed in different batches (same lab but at a different time\u0026ndash;typically a few days or weeks separation), and these can significantly impact the measurement of common bio-measures. Like many lab effects, batch effects can arise from differences in reagent lots, calibration standards, and instrument performance over time. These variations can introduce biases that obscure true biological signals, leading to inconsistent and unreliable biomarker readings. For example, in large epidemiological studies, blood samples might be collected over several months or years, and if processed in different batches, batch effects may distort the accurate biomeasurements such as lipid profiles or cytokines, ultimately confounding the study results.\u003c/p\u003e\u003cp\u003eThe influence of batch effects is particularly pronounced in high-throughput omics technologies where large datasets are generated (\u003cspan additionalcitationids=\"CR57 CR58 CR59\" citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e). In epigenetics, proteomics, and metabolomics studies, batch effects can drive significant discrepancies in biomarker quantification. These systematic differences can obscure biological variation and lead to erroneous conclusions. Strategies to mitigate these influences include randomized sample processing, the use of quality control samples across all batches, and the application of advanced statistical methods to adjust for batch variability. Properly accounting for batch effects is crucial to ensure robust and reproducible quantitative biomeasure analysis. Without rigorous quality control and validation, batch effects are difficult to detect\u003c/p\u003e\u003cp\u003eFor biomeasures that require long-term storage in the freezer, the freeze-thaw process can lead to significant measurement errors in the assessment of common bio-measures by causing physical and chemical alterations to biological samples (\u003cspan additionalcitationids=\"CR62 CR63\" citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e). Repeated freeze-thaw cycles can lead to the degradation of proteins, nucleic acids, and other biomolecules found in serum and plasma samples. This degradation results from mechanical stress and the formation of ice crystals, which can disrupt cell membranes and denature proteins. For instance, common biomarkers like cytokines and growth factors are particularly sensitive to these structural changes, leading to decreased concentrations and variability in measurement. Similarly, lipid components such as cholesterol and triglycerides can undergo hydrolysis and oxidation during freeze-thaw events, resulting in altered levels and compromised measurement accuracy. These changes are especially problematic for longitudinal studies where samples are stored and analyzed over extended periods.\u003c/p\u003e\u003cp\u003eAnother dimension of freeze-thaw cycle effects involves the stability of molecular biomarkers such as RNA and DNA. Nucleic acids are susceptible to degradation through repeated freezing and thawing, which can greatly affect the accuracy of PCR and sequencing-based biomarker assays (\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e). Ensuring minimal freeze-thaw cycles, along with optimized storage conditions, is critical to maintaining the fidelity of biomeasures, thus ensuring the reliability of clinical and research outcomes.\u003c/p\u003e\u003cp\u003eThe CLIA (Clinical Laboratory Improvement Amendments) program is specifically designed to set quality standards for laboratory testing conducted on human specimens, particularly for diagnostic purposes in healthcare settings across the U.S (\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e, \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e). To do this, laboratories must undergo a structured process to ensure they meet federal standards for clinical testing performed on human specimens. Based on the complexity of tests performed\u0026mdash;categorized as waived, moderate complexity, or high complexity\u0026mdash;the lab will receive a specific CLIA certification type (\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e). Laboratories performing moderate and high complexity tests must pay a fee, which varies according to the complexity level and the volume of tests performed. Labs performing moderate and high complexity testing, inspections are required to assess compliance with CLIA regulations. These labs must also participate in proficiency testing, which involves periodic external assessments to ensure the accuracy and reliability of their testing results. Waived test laboratories are generally not routinely inspected but must ensure they meet basic quality standards. While CLIA is a U.S.-based program, other countries have their regulatory frameworks and accreditation systems for clinical laboratories. For example, many countries recognize the ISO 15189 standard, which is an international accreditation for medical laboratories focusing on quality management and technical competence.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003e4.5 Raw Data to Edited Response: Processing Errors\u003c/h2\u003e\u003cp\u003eQuality Control (QC) metrics are crucial in minimizing measurement error in the assessment of common biomarkers by ensuring the reliability and reproducibility of laboratory results. QC procedures help in identifying and correcting systematic errors that may arise from instrument performance, reagent variability, and procedural inconsistencies. Through the regular evaluation of control samples with known concentrations, laboratories can monitor the accuracy and precision of the assays employed. For instance, in the measurement of glucose and cholesterol levels, QC metrics can detect and rectify drifts and shifts in the assay performance, thereby ensuring that subsequent measurements remain within acceptable error margins. The implementation of QC metrics not only helps in maintaining the consistency of test results within a single laboratory but also enhances comparability across different laboratories.\u003c/p\u003e\u003cp\u003eHowever, sometimes even small differences in advanced statistical QC methodologies have profound implications on measurement error in biomarker analysis (\u003cspan additionalcitationids=\"CR70\" citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e). For example, seemingly small, arbitrary differences in the quality control metrics for RNA can result in significantly different counts of transcripts\u0026ndash;yet most studies only apply one QC protocol with little examination of their effect. Yet for other biomarkers, the QC metrics are rigid, which can sometimes exclude more samples from some populations. Therefore, the integration of robust QC metrics into laboratory practices is indispensable for maintaining the integrity and utility of bio-measure data, but only if the QC protocols have been thoroughly examined in multiple populations.\u003c/p\u003e\u003cp\u003eTask-based cognitive tests are subject to different processing errors. In conventional parlance, \u0026ldquo;lab\u0026rdquo; signifies a brick-and-mortar place where physical specimens are received and processed. However, as detailed in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, biomeasures can take other forms \u0026ndash; cognitive, physical, and mental functioning, sensory, etc, and are subject to their own processing errors (\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e). Output from a sensor device, such as a high-precision accelerometer, for example, captures a person\u0026rsquo;s movement in digital form, and that raw data must be edited to create data for analysis. The editing process may require skills or tools the survey team does not possess, and the raw data must be transmitted to another organization for editing. In this case, the potential for processing error is analogous to that for physical specimens.\u003c/p\u003e\u003c/div\u003e"},{"header":"5. Trade-offs Between Types of Errors","content":"\u003cp\u003eTSE offers a framework for understanding the mechanisms that can lead to different types of errors. Investigating each data collection stage can help us better estimate and prevent potential issues that can decrease data quality. Another strength of the framework is highlighting the tradeoffs often involved in data collection and how different types of errors are connected. In biomeasure collection, one example of this is the tradeoff between sampling error (precision) and selection bias. Studies like the UK BioBank have large samples that can lead to an apparent increase in precision. Nevertheless, this comes at the cost of a likely selection bias, with significant groups being under-represented. Biomeasure collection in probability surveys focuses on minimising selection bias with a resulting loss in precision.\u003c/p\u003e\u003cp\u003eAnother tradeoff within survey biological data collection is between nonresponse and measurement. More detailed or precise measures, such as biomarkers measured from blood have, in principle, better measurement quality than alternative indicators but come with a decrease in sample size and potentially more selective samples. Collecting blood and extracting biomarkers from it entails explicit consent, blood draw, and complex processing of the blood. All of these can lead to a more selective sample.\u003c/p\u003e\u003cp\u003eYet another tradeoff that may appear between internal and external validity in choosing the location for specimen collection, physical or cognitive tests. Collecting data in a clinic ensures that the context and the machines used are standardised, increasing the measurement quality. Nevertheless, the artificial nature of the context can also decrease external validity. Similarly, collecting biomarkers in people's homes can be more realistic and align with their behaviour. Still, there may be more measurement errors due to differences in contextual factors and the lack of validation of instruments outside clinical contexts.\u003c/p\u003e\u003cp\u003eStudy designers and users of biosocial data must be aware of these tradeoffs. One strategy is to minimize the total error of the survey, focusing on the main analyses and questions that will be investigated. Unfortunately, estimating total error is often challenging without \u0026ldquo;gold standard\u0026rdquo; measures. An alternative approach is to minimise the errors most relevant to their research questions while accounting for possible unintended consequences on the other dimensions of the TSE. Researchers also need to account for costs (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). More intrusive biomeasures entail using staff with specialised skills and the costs of processing biological data collection. Additionally, collecting such data must account for the extra costs of processing and storing it. As a result, data collection is often a process of optimising data quality while keeping within budget constraints.\u003c/p\u003e"},{"header":"6. Conclusions","content":"\u003cp\u003eThis paper has explored the application of the Total Survey Error (TSE) framework to biomeasure data collection in social surveys, identifying the unique challenges and sources of error involved. The integration of biomeasures with survey data offers an invaluable new source of information, allowing researchers to investigate complex biosocial relationships. However, the process also introduces new error sources at various stages, from participant selection to data processing, necessitating an adaptation of the TSE framework. The updated TSE framework highlighted some new data collection stages and associated errors that researchers need to consider. On the measurement side, researchers need to be aware of the stage of specimen and raw data collection. These stages can lead to two new types of errors: transmission error and lab/batch effects. On the selection side, we also highlight two new steps: biomeasure response and raw data which can lead to two new types of errors: biomeasure nonresponse and data loss error. In reviewing the framework we also discussed how the other stages of the TSE framework can be influenced by biomeasure data collection.\u003c/p\u003e\u003cp\u003eA crucial point highlighted in this paper is the importance of using probability samples to make biosocial research better reflect the general population. By integrating biomeasures into probability-based social surveys, researchers can obtain more generalizable findings, which are vital for public health and policy-making. Developing best practices and sharing knowledge across the research community are essential for advancing the field of biosocial surveys. Unfortunately, much of the practical knowledge and methodological innovation in this area remains unpublished. It is vital to document and disseminate these insights to establish standardised protocols and improve data quality across studies.\u003c/p\u003e\u003cp\u003eFurther research is also needed to understand the effects of biomeasure collection on nonresponse. As collecting biomeasures often requires participants to undergo invasive procedures, this can impact response rates and introduce biases. Balancing the improved measurement accuracy from biomeasures with potential increases in nonresponse bias is essential for making informed decisions about study design. As many traditionally interviewer-administered surveys are moving towards self-completion, there is also a need to study the feasibility of collecting a broader range of biomeasures in self-administered surveys. A key question is which biomeasures respondents are willing to collect themselves, and whether such measures can be collected with the necessary precision and accuracy without the aid of a professional.\u003c/p\u003e\u003cp\u003eThe impact of interviewer effects on the collection of biomeasure data is another area requiring more investigation. Interviewers or nurses can add significant variability in the collection of biomeasure data, and there is increasing evidence that this phenomenon is widespread. More research is needed to understand how data collection agents impact substantive analyses. Additionally, investing in better training programs and developing standardised protocols can help minimize these effects and improve data reliability.\u003c/p\u003e\u003cp\u003eThere is also a significant gap in our understanding of how biomeasure data collection protocols need to be adapted for use outside of clinical settings. Collecting biomeasure data in diverse environments such as homes or mobile sites poses unique challenges in maintaining data integrity and consistency. More research is needed to develop appropriate protocols and scaling methods for these non-traditional settings.\u003c/p\u003e\u003cp\u003eAnother area that needs further development is understanding the relative sizes of the errors discussed in this paper. The TSE framework encourages us to consider multiple types of errors and the tradeoffs that appear during data collection. To inform data collection decisions, we need a better understanding of the magnitude of different types of errors and their relative importance.\u003c/p\u003e\u003cp\u003eFinally, the consistent use of terminology and frameworks like the TSE framework is crucial for accurately discussing and addressing the challenges in biomeasure data collection. Concepts such as target population and selection bias are fundamental for understanding data quality and should be uniformly applied across studies. The TSE framework provides a valuable tool for systematically addressing these issues, ensuring more valid, comparable results in biosocial research.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003cp\u003eNot applicable\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003cp\u003eAll authors have read and approved the manuscript and agree with its submission to BMC Medical Research Methodology. All authors meet authorship criteria, and all contributions and funding (if any) are disclosed in the manuscript.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eNot applicable\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eConceptualization and methodology (adapted Total Survey Error framework): A.C., B.E., C.M., J.W.S. B.E. wrote the first draft of the Introduction and the section on models for biomeasures. C.M. wrote the first drafts of the typology of biomeasures and the sections on laboratory and batch effects, transmission error, and processing errors. A.C. drafted the remaining sections, integrated contributions across authors, and coordinated the manuscript. All authors (A.C., B.E., C.M., J.W.S.) edited and revised the manuscript, approved the final version, and agree to be accountable for all aspects of the work. A.C. is the corresponding author.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eNot applicable\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eNot applicable\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eGroves RM, Lyberg L. Total Survey Error: Past, Present, and Future. Public Opin Q. 2010;74(5):849\u0026ndash;79.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHarris KM, Schorpp KM. Integrating biomarkers in social stratification and health research. Annu Rev Sociol. 2018;44(1):361\u0026ndash;86.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLenzenweger MF. Thinking clearly about the endophenotype\u0026ndash;intermediate phenotype\u0026ndash;biomarker distinctions in developmental psychopathology research. Dev Psychopathol. 2013;25(4pt2):1347\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMeloni M. How biology became social, and what it means for social theory. Sociol Rev. 2014;62(3):593\u0026ndash;614.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMeloni M. From boundary-work to boundary object: How biology left and re-entered the social sciences. Sociol Rev. 2016;64(1suppl):61\u0026ndash;78.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMitchell C, Schneper L, Notterman D, Faul J. Collection and Storage Effects on DNA Methylation and Telomere Length. 2024.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eQian W, Zhang C, Piersiak HA, Humphreys KL, Mitchell C. Biomarker adoption in developmental science: A data-driven modelling of trends from 90 biomarkers across 20 years. Infant Child Dev. 2024;33(1):e2366.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMitchell C, Schneper LM, Notterman DA. DNA methylation, early life environment, and health outcomes. Pediatr Res. 2016;79(1):212\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFreese J, Li JCA, Wade LD. The potential relevances of biology to social inquiry. Annu Rev Sociol. 2003;29(1):233\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJustice JN, Ferrucci L, Newman AB, Aroda VR, Bahnson JL, Divers J, et al. A framework for selection of blood-based biomarkers for geroscience-guided clinical trials. Geroscience. 2018;40:419\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKaplan JM, Wong HR. Biomarker discovery and development in pediatric critical care medicine. Pediatr Crit Care Med. 2011;12(2):165\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFalk EB, Hyde LW, Mitchell C, Faul J, Gonzalez R, Heitzeg MM, et al. What is a representative brain? Neuroscience meets population science. Proc Natl Acad Sci. 2013;110(44):17615\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMitchell C, McLanahan S, Brooks-Gunn J, Garfinkel I, Hobcraft J, Notterman D. Genetic Differential Sensitivity to Social Environments: Implications for Research. Am J Public Health. 2013;103(S1):S102\u0026ndash;10.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMayeux R. Biomarkers: Potential uses and limitations. NeuroRx. 2004;1:182\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSingh I, Rose N. Biomarkers in psychiatry. Nature. 2009;460(7252):202\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePragst F, Balikova MA. State of the art in hair analysis for detection of drug and alcohol abuse. Clin Chim Acta. 2006;370(1\u0026ndash;2):17\u0026ndash;49.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHollnagel E, editor. Handbook of Cognitive Task Design. CRC; 2003.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSakshaug JW, Ofstedal MB, Guyer H, Beebe TJ. The Collection of Biospecimens in Health Surveys. In: P. Johnson T, editor. Health Survey Methods [Internet]. 1st edn Wiley; 2014 [cited 2025 Sept 25]. pp. 383\u0026ndash;419. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://onlinelibrary.wiley.com/doi/\u003c/span\u003e\u003cspan address=\"https://onlinelibrary.wiley.com/doi/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/9781118594629.ch15\u003c/span\u003e\u003cspan address=\"10.1002/9781118594629.ch15\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBeebe TJ. Challenges to Collecting Survey-Based Biomarker and Genetic Data. Surv Res. 2007;(3).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAllicock OM, Petrone ME, Yolda-Carr D, Breban M, Walsh H, Watkins AE, et al. Evaluation of saliva self-collection devices for SARS-CoV-2 diagnostics. BMC Infect Dis. 2022;22(1):284.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNunes AP, Oliveira IO, Santos BR, Millech C, Silva LP, Gonz\u0026aacute;lez DA, et al. Quality of DNA extracted from saliva samples collected with the Oragene\u003csup\u003e\u0026trade;\u003c/sup\u003e DNA self-collection kit. BMC Med Res Methodol. 2012;12:1\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZarbl J, Eimer E, Gigg C, Bendzuck G, Korinth M, Elling-Audersch C, et al. Remote self-collection of capillary blood using upper arm devices for autoantibody analysis in patients with immune-mediated inflammatory rheumatic diseases. RMD Open. 2022;8(2):e002641.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcFall SL, Conolly A, Burton J. Collecting biomarkers using trained interviewers. Lessons learned from a pilot study. Surv Res Methods. 2014;8(1):57\u0026ndash;66.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHyland A, Ambrose BK, Conway KP, Borek N, Lambert E, Carusi C, et al. Design and methods of the Population Assessment of Tobacco and Health (PATH) Study. Tob Control. 2017;26(4):371\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZipf G, Chiappa M, Porter KS, Ostchega Y, Lewis BG, Dostal J. Health and Nutrition Examination Survey Plan and Operations, 1999\u0026ndash;2010. National Center for Health Statistics; 2013.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGroves RM. Three Eras of Survey Research. Public Opin Q. 2011;75(5):861\u0026ndash;71.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePashazadeh F, Cernat A, Sakshaug JW. The Effects of Biological Data Collection in Longitudinal Surveys on Subsequent Wave Cooperation. In: Advances in Longitudinal Survey Methodology [Internet]. John Wiley \u0026amp; Sons, Ltd; 2021 [cited 2024 Apr 9]. pp. 100\u0026ndash;21. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://onlinelibrary.wiley.com/doi/abs/\u003c/span\u003e\u003cspan address=\"https://onlinelibrary.wiley.com/doi/abs/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/9781119376965.ch5\u003c/span\u003e\u003cspan address=\"10.1002/9781119376965.ch5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBoyle P. others. Willingness to participate in health surveys involving biomarker data collection. J Aging Res. 2021;45(2):201\u0026ndash;18.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSakshaug JW, Couper MP, Ofstedal MB. Characteristics of physical measurement consent in a population-based survey of older adults. Med Care. 2010;48(1):64\u0026ndash;71.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCernat A, Sakshaug JW, Chandola T, Nazroo J, Shlomo N. Nurse Effects on Non-response in Survey-Based Biomeasures. Int J Soc Res Methodol 2021 July 4;24(4):487\u0026ndash;99.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGavrilova N, Lindau ST. Salivary Sex Hormone Measurement in a National, Population-Based Study of Older Adults. J Gerontol Ser B. 2009;64B(suppl1):i94\u0026ndash;105.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKumari M, Andrayas A, Al Baghal T, Burton J, Crossley TF, Jones KS, et al. A randomised study of nurse collected venous blood and self-collected dried blood spots for the assessment of cardiovascular risk factors in the Understanding Society Innovation Panel. Sci Rep. 2023;13(1):13008.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBenzeval M, Andrayas A, Mazza J. others. Does the feedback of blood results in observational studies influence response and consent? A randomised study of the Understanding Society Innovation Panel. BMC Med Res Methodol. 2023;23:134.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWest BT, Blom AG. Explaining Interviewer Effects: A Research Synthesis. J Surv Stat Methodol. 2017;5(2):175\u0026ndash;211.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCernat A, Sakshaug JW, Chandola T, Nazroo J, Shlomo N. Nurse Effects on Non-response in Survey-Based Biomeasures. Int J Soc Res Methodol. 2020;0(0):1\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eClarke P, Houle J. Selection bias in biomarker data: Using Heckman models to adjust for nonresponse in HIV prevalence estimates. Demography. 2014;51(2):975\u0026ndash;96.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHannemann T, Shlomo N, Chandola T, Chatzi G. Compensating for non-response in biosocial research: Simulation study from a cross-sectional analysis. 2018.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBossuyt PMM. Clinical validity: defining biomarker performance. Scand J Clin Lab Investig Suppl. 2010;242:46\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eM\u0026aacute;čov\u0026aacute; L, Bič\u0026iacute;kov\u0026aacute; M, Vitamin D. Current challenges between the laboratory and clinical practice. Nutrients. 2021;13(6):1758.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhan AI, Vasquez Y, Wians FH Jr, Kroll MH. The variability of results between point-of-care testing glucose meters and the central laboratory analyzer. Arch Pathol Lab Med. 2006;130(10):1527\u0026ndash;32.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMohamed R, Campbell JL, Cooper-White J, Dimeski G, Punyadeera C. The impact of saliva collection and processing methods on CRP, IgE, and Myoglobin immunoassays. Clin Transl Med. 2012;1:1\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMartin AR, Kanai M, Kamatani Y, Okada Y, Neale BM, Daly MJ. Clinical use of current polygenic risk scores may exacerbate health disparities. Nat Genet. 2019;51(4):584\u0026ndash;91.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWaldmann S, Sakshaug JW, Cernat A. Interviewer Effects on the Measurement of Physical Performance in a Cross-National Biosocial Survey. J Surv Stat Methodol. 2023;smad031.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCernat A, Sakshaug JW. The Impact of Nurse Continuity on Biosocial Survey Participation. Surv Methods Insights Field SMIF [Internet]. 2020 July 7 [cited 2020 July 7]; Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://surveyinsights.org/?p=13222\u003c/span\u003e\u003cspan address=\"https://surveyinsights.org/?p=13222\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCernat A, Sakshaug JW. Interviewer Effects in Biosocial Survey Measurements. Field Methods. 2021.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCrimmins EM, Zhang YS, Kim JK, Frochen S, Kang H, Shim H, et al. Dried blood spots: Effects of less than optimal collection, shipping time, heat, and humidity. Am J Hum Biol. 2020;32(5):e23390.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMurphy JM, Browne RW, Hill L, Bolelli GF, Abagnato C, Berrino F, et al. Effects of transportation and delay in processing on the stability of nutritional and metabolic biomarkers. Nutr Cancer. 2000;37(2):155\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu T, Rifai N, Roberts LJ, Willett WC, Rimm EB. Stability of measurements of biomarkers of oxidative stress in blood over 36 hours. Cancer Epidemiol Biomarkers Prev. 2004;13(8):1399\u0026ndash;402.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDakappagari N, Zhang H, Stephen L, Amaravadi L, Khan MU. Recommendations for clinical biomarker specimen preservation and stability assessments. Bioanalysis. 2017;9(8):643\u0026ndash;53.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcDade TW, Williams S, Snodgrass JJ. What a drop can do: dried blood spots as a minimally invasive method for integrating biomarkers into population-based research. Demography. 2007;44(4):899\u0026ndash;925.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLindrose AR, McLester-Davis LWY, Tristano RI, Kataria L, Gadalla SM, Eisenberg DTA, et al. Method comparison studies of telomere length measurement using qPCR approaches: A critical appraisal of the literature. PLoS ONE. 2021;16(1):e0245582.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu AHB. Biological and analytical variation of clinical biomarker testing: implications for biomarker-guided therapy. Curr Heart Fail Rep. 2013;10:434\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAylward LL, Hays SM, Koch HM, Jones KC, Levy L, Bevan R. Sources of variability in biomarker concentrations. J Toxicol Environ Health Part B. 2014;17(1):45\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eArch BN, Blair J, McKay A, Gregory JW, Newland P, Gamble C. Measurement of HbA1c in multicentre diabetes trials \u0026ndash; should blood samples be tested locally or sent to a central laboratory: an agreement analysis. Trials. 2016;17:8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTworoger SS, Hankinson SE. Use of biomarkers in epidemiologic studies: minimizing the influence of measurement error in the study design and analysis. Cancer Causes Control. 2006;17:889\u0026ndash;99.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHan W, Li L. Evaluating and minimizing batch effects in metabolomics. Mass Spectrom Rev. 2022;41(3):421\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTaioli E, Kinney P, Zeleniuch-Jacquotte A, Fulton H, Shore RE, Toniolo P, et al. Application of reliability models to studies of biomarker validation. Environ Health Perspect. 1994;102(3):306\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGoh WWB, Wang W, Wong L. Why batch effects matter in omics data, and how to avoid them. Trends Biotechnol. 2017;35(6):498\u0026ndash;507.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePrice EM, Robinson WP. Adjusting for batch effects in DNA methylation microarray data, a lesson learned. Front Genet. 2018;9:83.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLeek JT, Scharpf RB, Corrada Bravo H, Simcha D, Langmead B, Johnson WE, et al. Tackling the widespread and critical impact of batch effects in high-throughput data. Nat Rev Genet. 2010;11(10):733\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee JE, Kim SY, Shin SY. Effect of repeated freezing and thawing on biomarker stability in plasma and serum samples. Osong Public Health Res Perspect. 2015;6(6):357\u0026ndash;62.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRebholz SL, Melchior JT, Welge JA, Remaley AT, Davidson WS, Woollett LA. Effects of multiple freeze/thaw cycles on measurements of potential novel biomarkers associated with adverse pregnancy outcomes. J Clin Lab Med. 2017;2(1):10\u0026ndash;16966.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKeshavan A, Heslegrave A, Zetterberg H, Schott JM. Stability of blood-based biomarkers of Alzheimer\u0026rsquo;s disease over multiple freeze-thaw cycles. Alzheimers Dement Diagn Assess Dis Monit. 2018;10:448\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMitchell BL, Yasui Y, Li CI, Fitzpatrick AL, Lampe PD. Impact of freeze-thaw cycles and storage time on plasma samples used in mass spectrometry-based biomarker discovery projects. Cancer Inf. 2005;1:117693510500100110.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang C, Kohli N, Henn L. A Second-Order Longitudinal Model for Binary Outcomes: Item Response Theory Versus Structural Equation Modeling. Struct Equ Model Multidiscip J. 2015;0(0):1\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEhrmeyer SS, Laessig RH. Has compliance with CLIA requirements really improved quality in US clinical laboratories? Clin Chim Acta. 2004;346(1):37\u0026ndash;43.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNieto Sierra JA, Gefen D. 35 years after CLIA 1988: Key insights and policy implications among laboratory professionals. PLoS ONE. 2024;19(9):e0311251.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlexander TS. Licensure and Accreditation. In: Manual of Molecular and Clinical Laboratory Immunology. 2024. pp. 363\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHartley SW, Mullikin JC. QoRTs: a comprehensive toolset for quality control and data processing of RNA-Seq experiments. BMC Bioinformatics. 2015;16:1\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSheng Q, Vickers K, Zhao S, Wang J, Samuels DC, Koues O, et al. Multi-perspective quality control of Illumina RNA sequencing data analysis. Brief Funct Genomics. 2017;16(4):194\u0026ndash;204.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDrabovich AP, Mart\u0026iacute;nez-Morillo E, Diamandis EP. Toward an integrated pipeline for protein biomarker development. Biochim Biophys Acta BBA - Proteins Proteom. 2015;1854(6):677\u0026ndash;86.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRobinson P. Task\u0026ndash; based language learning: A review of issues. Lang Learn. 2011;61:1\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHerting MM, Gautam P, Chen Z, Mezher A, Vetter NC. Test-retest reliability of longitudinal task-based fMRI: Implications for developmental studies. Dev Cogn Neurosci. 2018;33:17\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-research-methodology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmrm","sideBox":"Learn more about [BMC Medical Research Methodology](http://bmcmedresmethodol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmrm/default.aspx","title":"BMC Medical Research Methodology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Total Survey Error, Biomeasures, Biosocial surveys, Data quality, Measurement error, Nonresponse bias, Laboratory effects, Survey methodology","lastPublishedDoi":"10.21203/rs.3.rs-7711723/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7711723/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003eThe integration of biomeasures into social surveys has grown rapidly, enabling new insights into biosocial processes by combining probability-based survey data with biological indicators. However, biomeasure collection outside clinical settings introduces unique sources of error that are not fully addressed by traditional survey methodology frameworks. This paper adapts the Total Survey Error (TSE) framework to account for the distinctive challenges of biomeasure data collection in surveys.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eWe review the stages of biomeasure data collection in social surveys and systematically map them onto the TSE framework. We extend the framework to identify new error sources, including biomeasure nonresponse, data loss during transmission, laboratory and batch effects, and interviewer-related biases. The discussion draws on evidence from large-scale biosocial surveys and methodological studies, highlighting trade-offs between different sources of error and the implications for study design, data processing, and analysis.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eOur extended framework demonstrates that biomeasure collection adds multiple new opportunities for error across both selection and measurement processes. On the selection side, additional nonresponse mechanisms\u0026mdash;such as refusal to consent to biological collection or ineligibility due to physical characteristics\u0026mdash;can create systematic biases. On the measurement side, issues such as interviewer variability, shipment and storage conditions, laboratory practices, and batch effects can significantly influence data quality. Trade-offs emerge between maximizing measurement precision and minimizing participation bias, between standardization in clinical contexts and ecological validity in household settings, and between cost constraints and optimal data integrity.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003eThe adapted TSE framework provides a structured approach to evaluating and minimizing errors in biomeasure data collection in social surveys. Recognizing and addressing new sources of error is essential for producing valid and generalizable biosocial research. Future work should focus on developing standardized protocols, improving interviewer and nurse training, expanding feasible self-collection methods, and quantifying the relative magnitude of different error sources. By systematically incorporating biomeasures into probability-based surveys while accounting for these challenges, researchers can strengthen the reliability and policy relevance of biosocial findings.\u003c/p\u003e","manuscriptTitle":"Total Survey Error in Biomeasure Data Collection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-10 12:41:10","doi":"10.21203/rs.3.rs-7711723/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"155764192611777074061895734170409311928","date":"2025-11-05T07:49:21+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"101436053071103511446370417336952988034","date":"2025-10-29T12:21:27+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-29T09:20:20+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-10-28T08:04:20+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-10-03T18:00:11+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-02T15:12:40+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Research Methodology","date":"2025-10-02T15:09:31+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-research-methodology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmrm","sideBox":"Learn more about [BMC Medical Research Methodology](http://bmcmedresmethodol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmrm/default.aspx","title":"BMC Medical Research Methodology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3bbe088b-f016-4753-8a88-490b3b7f19a9","owner":[],"postedDate":"November 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-11-10T12:41:10+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-10 12:41:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7711723","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7711723","identity":"rs-7711723","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}