Revisiting the Mental Health Impact of COVID-19 on Young Adults in the UK: Long-Term Trends, Temporary Setbacks, and Recovery

Revisiting the Mental Health Impact of COVID-19 on Young Adults in the UK: Long-Term Trends, Temporary Setbacks, and Recovery | 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 Revisiting the Mental Health Impact of COVID-19 on Young Adults in the UK: Long-Term Trends, Temporary Setbacks, and Recovery Golo Henseke, Ingrid Schoon This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6154489/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 09 May, 2025 Read the published version in Social Indicators Research → Version 1 posted You are reading this latest preprint version Abstract This study assesses the impact of the COVID-19 pandemic on the mental health of 16- to 29-year-olds in the United Kingdom, using longitudinal data from the UK Household Longitudinal Study (UKHLS) and its predecessor, covering the period from 2001 to 2023. The study identifies the causal effects of the lockdown (April 2020–March 2021) and the post-lockdown period (April 2021–March 2022) by estimating counterfactual mental health trajectories based on long-term trends. Unlike prior research, it accounts for potential reporting bias introduced by the UKHLS COVID-19 study. Mental ill-health among young adults had been rising for nearly two decades before the pandemic. During the lockdown period, the average General Health Questionnaire (GHQ-12) psychological distress score increased by 9% of its standard deviation, while the prevalence of clinically relevant psychological distress rose by 4.5 percentage points. This impact was temporary, with mental health levels returning to predicted trends by April 2021, suggesting no lasting 'scar' on average mental health. The recovery coincided with declining feelings of loneliness and increased life satisfaction. The study also identifies variations in the pandemic’s mental health effects by gender, household income, age, and ethnicity. Women and young adults in the top third of the household income distribution experienced a more pronounced increase in psychological distress during lockdown. However, there is no evidence that the under-30 age group suffered, on average, more severe mental health effects than the rest of the adult population under 60 during the lockdown period. The findings challenge prevalent narratives by demonstrating the relative resilience of young adults in the face of the pandemic. Epidemiology Health Economics & Outcomes Research Psychological distress COVID-19 Young Adults Longitudinal Study Survey Design Effects Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction This paper assesses the mental health of 16–29-year-olds in the United Kingdom between 2001 and 2023 before, during, and after the COVID-19 pandemic, focusing on the causal effects of the lockdown and post-lockdown period (Hernán, 2018 ). It analyses long-term trends in psychological distress and evaluates the pandemic’s impact during the lockdown and post-lockdown period, accounting for differences in self-reported mental health across UK Household Longitudinal Study (UKHLS) surveys. From the outset, there were major concerns about the far-reaching mental health effects of COVID-19 (Hotopf et al., 2020 ). For example, Ahmed et al. ( 2023 ) counted 177 studies on pandemic-related mental health outcomes in Europe alone. While young people were less vulnerable to the disease’s direct health risks, stringent containment measures disrupted their education, careers, leisure activities, and peer relationships, potentially affecting their developmental trajectories (Settersten et al., 2020 ). Numerous studies reported a pronounced deterioration in mental health during the pandemic’s early phase in the UK (Fancourt et al., 2021 ), with young people and women disproportionately affected (Banks & Xu, 2020 ). However, research had to rely on convenience samples (O'Connor et al., 2021 ), did not adjust for pre-existing trends in mental health (Niedzwiedz et al., 2021 ), or did not consider how differences between surveys might have influenced self-reported mental health (Anaya et al., 2023 ; Pierce et al., 2020 ), potentially biasing estimates of pandemic-related impacts. Indeed, longitudinal evidence outside the UK represents a less clear-cut picture of elevated mental health risks (e.g., Jaschke et al., 2023 ). More recent global reviews of high-quality longitudinal evidence suggest that mental health remained unchanged or worsened only minimally during the pandemic—even among adolescents and young adults (Sun et al., 2023 ). These findings contrast with narratives of a widespread and lasting decline in population well-being. Moreover, measurement issues have received relatively little attention, even though many UK-based studies rely on pre- and post-pandemic comparisons using different surveys. Furthermore, evidence on mental health recovery following the vaccine rollouts, the lifting of lockdown restrictions, and economic rebounds remain limited. While receding stressors might be expected to improve young adults’ psychological well-being, prolonged and repeated disruptions could have developed lasting ‘scarring’ effects on psychosocial development. Existing research presents mixed findings, reflecting context-specific mental health trajectories and differences in how well-being is conceptualised. For example, in the UK, Gagné et al. ( 2022 ) documented a sharp increase in the risk of psychological distress (GHQ-12) between 2018/19 and April 2020, followed by a decline between July and September 2020 in individuals aged 16–34. Similarly, Henseke et al. ( 2022 ) observed a substantial rise in life satisfaction between February 2021 and May 2022, linked to increased social interactions and reduced uncertainties about learning and careers. The vaccine rollout also improved mental health, particularly among clinically vulnerable groups (Chaudhuri & Howley, 2022 ). Feelings of loneliness had returned to pre-pandemic levels by September 2021 (Kung et al., 2023 ). In Norway, Kozák et al. ( 2023 ) report above-average scores of depressive symptoms among adolescents in 2021 but not in 2022, suggesting a recovery of mental health to long-term trends. In contrast, Neugebauer et al. ( 2023 ) reported persistent declines in life satisfaction in a sample of German high school students, and Dhensa-Kahlon et al. ( 2025 ) found elevated psychological distress in UK adults aged 18–29 post-lockdown, though the extent to which these changes were pandemic-driven remains unclear. This contrasting evidence underscores the importance of contextualising pandemic-related mental health trends within pre-existing dynamics. A growing body of research indicates that young people’s mental health has been declining since at least the 2008 Great Recession (Blanchflower, Bryson, et al., 2024 ), as reflected in lower life satisfaction (Gagné et al., 2022 ), more reported symptoms of distress (Zhang et al., 2023 ), and rising self-harm (McManus et al., 2019 ). Irrespective of the causes, this ongoing deterioration suggests that distress levels in 2020 and beyond may have changed due to pre-pandemic trends rather than the pandemic itself. Additionally, psychological well-being tends to decline during the transition to adulthood (Blanchflower, Bryson, & Xu, 2024 ), explaining part of the mental health trends observed during the pandemic (Wright et al., 2024 ). Finally, small changes in survey design and administration can influence well-being estimates (Blanchflower, 2025 ; Conti & Pudney, 2011 ; Davillas et al., 2023 ), complicating longitudinal comparisons across studies. Using longitudinal survey data from the UK Household Longitudinal Study (UKHLS, also known as Understanding Society), this study makes several key contributions. Firstly, we compare self-reported mental health between the UKHLS main survey and the COVID-19 study to evaluate whether questionnaire design influenced reported distress levels. Secondly, we track changes in psychological distress among 16–29-year-olds since 2001, adjusting for individual fixed effects (including cohort effects) and age-related life course changes (e.g., employment and household composition). Thirdly, we use a Neyman-Rubin causal approach to identify the average treatment effect on the treated of the COVID-19 pandemic, estimating individual-specific counterfactual levels of distress had 2020 and 2021 been ‘normal’ years, while accounting for differences between surveys. This study aligns with prior UKHLS-based research on pandemic-related mental health impacts. Pierce et al. ( 2020 ) seminal contribution assessed the change in mental health in April 2020 against a counterfactual prediction of what would have been expected from population trends up to 2018/2019, finding a sharp increase in GHQ-12 distress among individuals under 35. However, their estimates did not account for age-specific mental health trends. Banks and Xu ( 2020 ) adjusted for these trends, confirming disproportionate mental health impacts on young adults and women. Gagné et al. ( 2021 ) showed that distress declined unevenly after April 2020, with slower recovery among women and young adults aged 25–34. More recently, Anaya et al. ( 2023 ) used a difference-in-differences approach, finding a + 32% standard deviation increase in GHQ-12 scores among 18–34-year-olds during the first national lockdown. Serrano-Alarcón et al. ( 2022 ) identified a reduction in mental ill-health from the easing of containment measures, pointing towards greater responsiveness of mental health to the behavioural restrictions than to the virus itself. Duarte Neves et al. ( 2024 ) also deployed a difference-in-difference approach, finding evidence for long-term ‘scarring’ of mental health from the pandemic for some population groups. All these studies combined data collected through a shortened web questionnaire launched in April 2020 with the UKHLS main survey, assuming rather than assessing measurement equivalence in self-reported mental health problems between the surveys. The remainder of the paper is structured as follows. The next section introduces the datasets, measures, and the analytical strategy. Section 3 presents and discusses the findings. The final section concludes. Methods Datasets The analysis draws on combined survey data from the first 14 waves of the UK Household Longitudinal Study (UKHLS), waves 11–18 of its predecessor, the British Household Panel Study (BHPS), and all sweeps of the UKHLS COVID-19 web survey (2020–2021). All data files are available for research at https://doi.org/10.5255/UKDA-Series-2000053 . UKHLS is an ongoing panel survey of about 40,000 households in the United Kingdom, launched in 2009 as a continuation of the BHPS, which ran from 1991 to 2008 (Institute for Social and Economic Research, 2023 ). The most recent wave (Wave 14) includes 35,500 individuals, nearly 6,000 of whom are under 30 years old. Fieldwork for each wave spans three years, with samples continually issued during the first two years. For example, UKHLS Wave 1 ran from 2009 to 2011 and Wave 14 issued samples from 2021 to 2023 1 , creating overlapping fieldwork periods in each calendar year. Adults aged 16 and over in sampled households are re-interviewed annually, including core members (initial sample members and their descendants) who move or form a new household. The BHPS followed a similar panel study protocol but with notable differences. It was smaller in scale, with full UK-wide coverage starting in 2001/2002 (Wave 11). Fieldwork typically began in September and lasted until April the following year. The final BHPS wave (2008/2009) collected data from 14,400 individuals (3,200 under 30 years old) across 8,100 households. UKHLS continued the BHPS sample from Wave 2 onward. Due to lockdown restrictions, the UKHLS main survey shifted to a web-first mode in mid-March 2020, with most responses collected online and a smaller proportion via telephone. Face-to-face fieldwork commenced in April 2022. However, even before the pandemic, 70% of panel members were already invited to complete the questionnaire web-first (Burton et al., 2020 ). This means that the transition away from in-person interviews in 2020 was an acceleration of an existing trend rather than an abrupt shift. Starting in April 2020, UKHLS participants were invited to complete a 20-minute web survey on their experiences and reactions to the COVID-19 pandemic (Institute for Social and Economic Research, 2021b ). The first four waves of the COVID-19 study were conducted monthly between April and July 2020, followed by bi-monthly surveys from September 2020 to March 2021 (waves 5–8). A final ninth wave was conducted in September 2021. As with the main survey over this period, responses were collected predominantly online, with a minor telephone mode available in waves 2 and 6. In the under-30 age group, telephone responses accounted for 39 cases in total (0.3% of all COVID-19 study cases under 30), which we integrated into the web survey dataset and did not treat separately. Response rates to the COVID-19 study varied from 42% in wave 1 to 29% in waves 6 and 7 (Institute for Social and Economic Research, 2021a ). The COVID-19 study is fully integrated with UKHLS, using shared panel IDs and including core demographic, life course, and mental health information. We extracted mental health, demographics, and life course data for 16–29-year-olds across UKHLS and BHPS survey waves since 2001. After removing singleton observations, there were 120,987 person-year observations from 22,247 individuals to measure long-term changes in mental health from 2001–2022 in the age-group 16–29 years. To assess the mental health effects of COVID-19, we restrict the sample to cases aged 16–29 years in the UKHLS waves 1–14 (2009–2023) and the COVID-19 study (2020–2021). We employ a complete-case analysis in the unbalanced panel, acknowledging that sample attrition and excluding respondents with missing data may introduce bias. The limitations of this approach and robustness checks assessing attrition bias will be discussed in subsequent sections, where we introduce an added-variable test to examine whether panel retention is systematically associated with mental health outcomes in our estimation model. In all, the COVID-19 sample consists of 96,564 person-wave observations from 18,027 individuals aged 16–29, with an average panel retention of 5.4 waves per respondent. Figure 1 shows the distribution of observations over calendar quarters (2020–2023) and their source. The COVID-19 web surveys contributed the bulk of cases over the pandemic, at 64% in 2020 and 43% in 2021, while UKHLS main survey data provided a consistent reference point for benchmarking changes in reported mental ill-health. To compare patterns with older age groups, some analyses will lift the age restriction to include cases up to age 59 years. Measures Mental health is assessed using the 12-item General Health Questionnaire (GHQ-12), a widely used, validated, and reliable instrument for measuring non-specific psychological distress in longitudinal samples (Lundin et al., 2016 ) and youth populations (Baksheev et al., 2011 ). The GHQ-12 has been administered in every wave of the UKHLS, its COVID-19 study and BHPS, ensuring consistent measurement over time. Moreover, prior research has found no evidence of panel conditioning effects in GHQ-12 responses (Pevalin, 2000 ), further supporting the instrument’s consistency over repeated administrations. For the analysis, two complementary measures of psychological distress are used, both derived from the same GHQ-12 instrument. The first is a continuous GHQ-12 score, constructed by summing responses across the twelve items, where each item is scored on a four-point Likert scale: 0 for “not at all,” 1 for “no more than usual,” 2 for “rather more than usual,” and 3 for “much more than usual.” The total GHQ-12 score ranges from 0 to 36, with higher values indicating greater psychological distress. The second measure is a binary indicator of psychological distress derived from the GHQ-12 instrument, identifying individuals experiencing clinically relevant distress if they reported “rather more than usual” or “much more than usual” on at least four of the twelve items. This cutpoint-based classification, commonly used in epidemiological and social science research (Pierce et al., 2020 ), helps to distinguish clinically relevant cases of psychological distress. Figure 2 shows the distribution of the GHQ-12 sum score in the trend sample, demonstrating that the measure effectively captures a broad range of distress levels. The distribution is right-skewed, with most observations concentrated in the lower-to-mid range but with a clear spread toward higher distress levels. Importantly, there is no apparent clustering near the upper end of the scale, confirming that the measure retains sensitivity at higher distress levels. This pattern suggests that ceiling effects are unlikely to be a major concern, as the GHQ-12 scale allows room for capturing worsening psychological distress if present. In addition to mental health measures, for subgroup analyses and to control for life course milestones, we extract information on individual age, sex, ethnic minority status, employment status, partnership status, and the number of young children under five in the household. Socioeconomic status is measured using the percentile household income rank at the time individuals entered the panel, based on a comparison of gross monthly household income equivalised for household composition using the modified OECD scale, ranked within each survey wave (Anyaegbu, 2010 ). To measure time trends and survey effects, interview date information is incorporated to adjust for long-term changes and seasonal fluctuations in mental health outcomes. A binary indicator for the COVID-19 web study differentiates between the main survey and COVID-19 study observations. Table 1 presents summary statistics for both the long-term trend sample (2001–2023) and the UKHLS sample (2009–2023). The two samples are highly comparable in terms of demographic composition and key characteristics. The mean GHQ-12 score was 11.4 in the trend sample and 11.6 in the UKHLS sample, with a similar proportion classified as psychologically distressed (23% vs. 24%). Both samples were skewed towards women (57% in the trend sample and 58% in the UKHLS sample), and about one-quarter of respondents identified as belonging to an ethnic minority group (24% vs. 27%). The distribution of age groups was stable across samples, with approximately 30% aged 16–19, 35% aged 20–24, and 35% aged 25–29. Socioeconomic factors also showed minimal differences; the mean household income rank was 49.41 in the trend sample and 48.20 in the UKHLS sample. Similarly, employment rates (54% vs. 53%) and partnership status (27% vs. 26%) were nearly identical. Table 1 Summary Statistics for Samples of 16-29-year-olds Variable N Individuals Mean Std. Dev Std. Dev (within) Long-Term Trend Sample 2001–2023 GHQ-12 (Score) 120987 22247 11.41 5.95 4.17 GHQ-12 (Cases) 120987 22247 0.23 0.42 0.32 Female 120984 22246 0.57 0.50 0.00 Age: 16–19 120987 22247 0.30 0.46 0.33 Age: 20–24 120987 22247 0.35 0.48 0.40 Age: 25–29 120987 22247 0.35 0.48 0.33 Ethnic minority 119012 21777 0.24 0.42 0.00 Household income rank 119632 21973 49.41 27.84 0.00 In work 120987 22247 0.54 0.50 0.35 Living as a couple 120987 22247 0.27 0.45 0.24 Number of children < 5 in the household 120987 22247 0.18 0.46 0.27 UKHLS Sample 2009–2023 GHQ-12 (Score) 96564 18027 11.64 6.03 4.20 GHQ-12 (Cases) 96564 18027 0.24 0.42 0.33 Female 96561 18026 0.58 0.49 0.00 Age: 16–19 96564 18027 0.30 0.46 0.32 Age: 20–24 96564 18027 0.36 0.48 0.40 Age: 25–29 96564 18027 0.35 0.48 0.32 Ethnic minority 95044 17659 0.27 0.44 0.00 Household income rank 95222 17757 48.20 27.68 0.00 In work 96564 18027 0.53 0.50 0.35 Living as a couple 96564 18027 0.26 0.44 0.23 Number of children < 5 in the household 96564 18027 0.17 0.45 0.25 Source: UKHLS and BHPS main surveys, UKHLS COVID-19 web surveys. Sample of 16-29-year-olds. Authors’ calculations. Analytical Approach Our analytical approach estimates the causal impact of the COVID-19 pandemic on young adults’ mental health by comparing observed outcomes during the pandemic with counterfactual outcomes predicted from pre-pandemic trends. We begin with a general model of individual mental health over time: $$\:M{H}_{it}={\alpha\:}_{it}+{\gamma\:}_{t}\:COVID\left(t\right)+{\gamma\:}_{s}\:COVID\left(s\right)+{ϵ}_{it}$$ Where: \(\:M{H}_{it}\) represents the mental health outcome (GHQ-12 scores or cases indicator) for individual \(\:i\:\) at time \(\:t\) . \(\:{\alpha\:}_{it}\) is an individual-specific, time-dependent component that reflects life-course variation in predisposition towards mental (ill-)health. \(\:{\gamma\:}_{t,s}\:\) captures the effect of the COVID-19 pandemic on mental health, distinguishing between the lockdown period (t = 1) and post-lockdown period (s = 1). \(\:{ϵ}_{it}\) represents idiosyncratic mental health shocks. Our ‘estimand’ or target parameter is the Average Treatment Effect on the Treated (ATT) for the lockdown and post-lockdown periods (Lundberg et al., 2021 ). Specifically, for the lockdown period we define $$\:AT{T}_{t}=E\left[M{H}_{it}^{1}-{MH}_{it}^{0}|COVID\left(t\right)=1\right]$$ with an analogous definition for the post-lockdown period. This means the pandemic effect is conceptualised as the difference between the observed average mental health outcomes and the counterfactual scenario where the pandemic did not occur. To estimate the ATT, we employ a linear fixed effects regression model that accounts for individual-specific time-invariant heterogeneity and flexible time trends. The model is specified as: $$\:M{H}_{it}={\alpha\:}_{i}+g\left(t\right)+{X}_{it}\beta\:+{{\delta\:}_{t}D}_{it}+{\gamma\:}_{t}\:COVID\left(t\right)+{\gamma\:}_{s}\:COVID\left(s\right)+{\epsilon\:}_{it}$$ 1 Where: \(\:{\alpha\:}_{i}\) captures individual-specific fixed effects. g(t) is a function of time that captures period trends in mental health. Initially, we model g(t) as a cubic polynomial in survey years to capture long-term trends since 2009, following prior research (Banks & Xu, 2020 ; Pierce et al., 2020 ). We later test the robustness by replacing the cubic polynomial with a restricted cubic spline in interview dates. \(\:{X}_{it}\) is a vector of time-varying covariates, including age bands (< 20, 20–24, 25+), partnership status, the number of children under 5 in the household, and survey month. These covariates serve as proxies for life-cycle markers, helping to approximate age-related effects and adjust for seasonal fluctuations in mental health. \(\:{D}_{it}\:\) time-varying survey design effects (dummy variable for COVID-19 surveys vs UKHLS main survey), accounting for potential level differences in self-reported mental health across surveys. \(\:COVID(t,s)\) are indicator variables for the lockdown and post-lockdown periods, with the corresponding parameter \(\:{\gamma\:}_{t,s}\) capturing the pandemic’s effects on mental health. \(\:{\epsilon\:}_{it}\) is the error term, assumed to be independent of the covariates, time trend and individual-fixed effects. Under the assumption that our empirical model is correctly specified—meaning that individual fixed effects, time trends, and covariates fully capture the counterfactual mental health trajectory in the absence of the pandemic—the coefficients \(\:{\gamma\:}_{t,s}\) in Eq. ( 1 ) represent the causal impact of the COVID-19 pandemic on mental health. Formally, we have $$\:AT{T}_{t,s}=E\left[M{H}_{i}^{1}\left(t,s\right)-{MH}_{i}^{0}\left(t,s\right)\right|COVID(t,s)=1]={\gamma\:}_{t,s}\:$$ which rests on the assumption that the counterfactual mental health —what would have occurred absent the pandemic—is fully captured by the pre-pandemic trends, individual fixed effects, and covariates: \(\:{\alpha\:}_{i}+g\left(t\right)+{\varvec{X}}_{it}\beta\:+{{\delta\:}_{t}D}_{it}\) . This approach relies on the parallel trends assumption , which posits that, in the absence of COVID-19, mental health outcomes during the pandemic would have followed the same underlying trajectory as in the pre-pandemic period. By incorporating individual fixed effects and a flexible time trend, our model effectively imposes this parallel trend assumption. Therefore, if this assumption holds, then any estimated deviations from the expected trajectory during the pandemic—captured by the COVID-19 period indicators—reflect the ATT. To strengthen confidence in the parallel trends assumption and address potential threats to identification, we conduct several supplementary analyses and robustness checks: 1. Covariate Adjustment : While the model assumes that COVID-19’s effects on mental health do not operate through life course markers, this assumption can be contested. We assess sensitivity by estimating the model without these covariates and comparing results. 2. Event Study : To examine the evolution of mental health changes over shorter intervals, we break the lockdown period into calendar quarters, allowing us to verify whether pre-pandemic trends closely mirror the dynamics observed during the pandemic in the absence of treatment. 3. Subgroup Analysis : We conducted subgroup analyses to examine whether the pandemic’s effects varied across different demographic and socioeconomic groups. This helps illuminate the stability of our estimated effect, adding substantive insights as well. 4. Alternative Measures of Subjective Well-Being : We re-estimate models using alternative measures of subjective well-being (single-item life satisfaction and feelings of loneliness), testing whether findings are robust to different operationalisations of mental health and, thus, potentially different measurement errors. 5. Attrition Bias Testing : Given that panel dropout could be related to mental health, we explicitly test for attrition bias to ensure that selection effects do not confound estimates. 6. Alternative Time Trend : We replaced the cubic polynomial in survey years with a restricted cubic spline in interview dates to assess robustness against different specifications of the long-term time trend. 7. Placebo Test : Finally, we conducted a placebo test, assigning pseudo-treatment periods to years before the pandemic. The absence of significant effects in these placebo tests reinforces the credibility of our identification strategy. Together, these checks address potential threats to identification and provide a framework for interpreting our estimates of the lockdown and post-lockdown effects as the causal impact of the COVID-19 pandemic and associated restrictions on mental health. Additionally, these analyses offer substantive insights into how young people coped with the pandemic and its immediate aftermath. While further discussion of life-cycle (age, period, and cohort) effects and the role of specific covariates is provided in the findings section, it is worth noting here that controlling for partnership status, the presence of young children, and broad age groups along with survey month adjustments, helps to separate period-specific effects from broader life-course dynamics. All data cleaning, management, and analyses were conducted in Stata 18.5. Syntax files to replicate analysis, tables and figures are hosted at https://doi.org/10.5522/04/28469006.v1 . Findings Differences in Psychological Distress Between the UKHLS Main Survey and the COVID-19 Study Pooling observations from the UKHLS main survey and the COVID-19 study assumes that mental health measures are comparable across surveys. This assumption is based on the use of the same GHQ-12 instrument and the fact that both surveys were administered primarily via web questionnaires during the pandemic. However, if the questionnaire design introduced systematic measurement differences, pooling the two data sources without adjustment may introduce bias in longitudinal analyses of mental health before, during, and after the pandemic. To test for potential discrepancies, this section compares average levels of the GHQ-12 sum score and the prevalence of clinically relevant cases of psychological distress between the main survey and the COVID-19 study conducted in the same survey year-month. Table 2 presents the results, comparing average differences in reported psychological distress between surveys in the age group 16–29 with figures for 30–44 and 45-59-year-olds. The findings suggest statistically significant differences in reported mental health levels between the two surveys, particularly among younger adults. In the 16–29 age group, the COVID-19 study recorded higher distress levels than the main survey, with an average GHQ-12 score difference of 0.8 points (95% CI [0.46, 1.14]) and a 2.1 percentage point increase in the prevalence of psychological distress (95% CI [-0.002, 0.044]), although the latter was not statistically significant at conventional levels. For individuals aged 30–44, the COVID-19 study also reported higher distress levels, with an average GHQ-12 score difference of 0.59 points (95% CI [0.32, 0.86]) and a 2.1 percentage point increase in the prevalence of psychological distress (p < 0.05). In contrast, there is no statistical evidence for survey effects on GHQ-12 scores among respondents aged 45 and above. Table 2 Average Differences in Reported Mental Health Problems Between the UKHLS Main Survey and the COVID-19 Study by Age Group (N = 175,406). Age Group GHQ-12 (Score) GHQ-12 (Cases) 16–29 0.796 *** (0.173) 0.021 (0.012) 30–44 0.588 *** (0.137) 0.021 * (0.009) 45–59 0.072 (0.103) -0.009 (0.007) Note: For each outcome, a survey-weighted regression was estimated with interactions among gender, age group, and COVID-19 study status, adjusting for the year and month of sample issuance. The average marginal effects of the COVID-19 study were computed by age group. Standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001 Source: UKHLS main waves 12 and 13, UKHLS COVID-19 study. Author’s calculations. These findings suggest that reported levels of psychological distress differ between the UKHLS main survey and the UKHLS COVID-19 study, with elevated distress among younger respondents and lower distress among older respondents in the COVID-19 study relative to the main survey. The observed differences could be driven by selection effects, survey design variations, or administrative differences. While these results do not imply that one survey provides a more accurate estimate of mental health, they underscore the importance of accounting for potential measurement differences. Failing to do so could lead to biased estimates when assessing changes in mental health before, during, and after the pandemic, particularly for younger age groups. Long-Term Trends in Mental Health Since 2001 This section examines long-term trends in mental health among 16-29-year-olds from 2001 onward, based on linear fixed effects estimations of Eq. ( 1 ). At this stage, the focus is on the underlying time trend \(\:g\left(t\right)\) , which is modelled using year dummies. To isolate period trends from confounding life course effects, the model includes individual fixed effects, which adjust for cohort differences and three broad age groups (< 20, 20–24, and 25+), effectively constraining age effects. This effectively transforms Eq. ( 1 ) into an age-period-cohort (APC) model (Fosse & Winship, 2019 ). Figure 3 displays the estimated change in psychological distress over time, adjusted for individual heterogeneity and life stage. The estimates indicate a steady and substantial increase in psychological distress among young adults. Between 2001 and 2019, the GHQ-12 score increased by 2.9 score points within individuals on average (Fig. 3 , left plot), a difference that is both statistically and quantitatively significant. This increase represents 49% of the pooled standard deviation of the GHQ-12 scale (95% CI [41.5%, 57.5%]). A similar trend is observed in the proportion of clinically relevant cases of psychological distress (Fig. 3 , right panel). Compared to 2001, its prevalence had risen by 8.9 percentage points within individuals by 2019. In relative terms, young adults in 2019 were 39% more likely to meet the distress threshold compared to those in 2001 (95% CI: [18.5%, 59.3%]), an increase equivalent to approximately 1.1 million additional cases per year in the age group. Both plots peaked in 2020, followed by improvements, especially in cases of psychological distress. The following section will delve deeper into these patterns to unpick what changes were due to COVID-19 and the result of pre-existing trends. For context, the trend towards mental ill-health, with a subsequent flatting in the wake of the COVID-19 pandemic, correlated with averages in prescribed medication in mental health per 15–29-year-old per year in England since 2016 (GHQ-12 Score: r = 0.96, p < 0.001; GHQ-12 Cases: r = 0.79, p = 0.020). The long-term increase in psychological distress among young adults highlights the need to assess potential pandemic effects conditional on changes that might have happened irrespective of the pandemic. Ignoring the long-term trends towards mental ill-health can upward bias or overstate the estimated impact of the COVID-19 shock on mental health and understate the subsequent recovery. COVID-19 Impact: Deterioration and Recovery Table 3 presents the estimated impact of the COVID-19 pandemic on psychological distress, based on Eq. ( 1 ). Column (1) provides estimates from a baseline model that adjusts for time trends but does not account for survey differences between the UKHLS main survey and the COVID-19 study. Column (2) presents the headline findings, incorporating covariates to adjust for survey design effects. Column (3) restricts the analysis to UKHLS main survey data, allowing for an assessment of how well the dummy variable adjustment strategy accounts for survey differences. Column (4) examines the sensitivity of the headline results by excluding life-course markers from the model. The top panel of Table 3 reports results for the GHQ-12 score, while the bottom panel presents findings for GHQ-12 cases with clinically relevant levels of psychological distress. Column (1) estimates a 0.95-point increase in the GHQ-12 mean score and a 6 percentage-point increase in clinically relevant cases of psychological distress during the lockdown period relative to the expected trend. Levels of psychological distress returned to trend in the post-lockdown period. However, these estimates do not account for potential survey mode effects. Column (2) presents the headline findings, which incorporate adjustments for survey differences. The estimates represent the difference between observed mental ill-health and the counterfactual level of mental ill-health had pre-pandemic trends continued and in the absence of survey response effects. The GHQ-12 mean score was 0.54 points higher than expected during the lockdown, while post-lockdown levels were 0.15 points below expected trends, although this difference was not statistically significant. Table 3 Mental Health During and After the COVID-19 Pandemic. (1) (2) (3) (4) Trend adjusted + Design adjusted Mainstage Only W/o Covariates GHQ-12 Score Lockdown 0.951 *** (0.075) 0.538 *** (0.080) 0.541 *** (0.080) 0.540 *** (0.080) Post-Lockdown 0.030 (0.087) -0.150 (0.090) -0.139 (0.091) -0.150 (0.090) Difference -0.921 *** (0.089) -0.688 *** (0.099) -0.680 *** (0.100) -0.690 *** (0.099) GHQ-12 Cases Lockdown 0.060 *** (0.006) 0.045 *** (0.006) 0.046 *** (0.006) 0.045 *** (0.006) Post-Lockdown -0.006 (0.007) -0.013 (0.007) -0.012 (0.007) -0.013 (0.007) Difference -0.067 *** (0.007) -0.058 *** (0.008) -0.058 *** (0.008) -0.058 *** (0.008) Seasonally adjusted X X X X Life course controls X X X Design X X X Time trend X X X X Observations 96,564 96,564 84,745 96,564 Individuals 18,027 18,027 17,985 18,027 Note: Results from linear fixed effects regression models estimating Eq. ( 1 ) for GHQ-12 scores and GHQ-12 cases. All models control for individual fixed effects, interview month, and a cubic polynomial in survey years since 2009. Column (2) adjusts for survey design effects using dummy variables for the COVID-19 study, interacted with interview year-quarter dummies. Column (3) restricts the sample to UKHLS main survey data. Column (4) excludes life-course markers (employment, partnership status, number of children under five). Standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001 Source: UKHLS, UKHLS COVID-19 survey. Authors’ calculations. The increase in GHQ-12 scores during the lockdown period, while statistically significant, was quantitatively modest, amounting to 9.2% of the score’s standard deviation (95% CI: [6.5%, 11.8%]). Similarly, the prevalence of clinically relevant psychological distress was 4.5 percentage points above expected levels in the lockdown phase, implying that young adults were 19% more likely to meet the distress threshold (95% CI: [13.9%, 24.6%]). After March 2021, psychological distress recovered to expected levels, consistent with a return to the long-term trend. Taken together, these findings indicate that the direct impact of the COVID-19 pandemic on psychological distress was moderate and time-limited for the average 16-29-year-old with no apparent scarring of mental health from repeated lockdowns. The initial shock was followed by a recovery to trend levels. Column (3) removes observations from the COVID-19 study, leaving only UKHLS main survey data. The similarity between columns (2) and (3) suggests that the dummy variable adjustment approach was sufficient to adjust for survey response effects on mental health reporting. Column (4) removes life-course markers from the model, with minimal impact on the headline findings. This suggests that the pandemic effect did not operate through partnership and family formation. Once long-term trends and survey response effects are accounted for, the findings suggest that the mental health impact of the COVID-19 shock among young adults was limited in duration, with levels returning to expected trends after March 2021. Despite the substantial disruptions experienced during the pandemic, these findings indicate that, on average, young adults did not experience lasting scarring effects. However, this temporary setback and subsequent recovery must be understood within the broader context of a sustained long-term rise in psychological distress, which preceded the pandemic. Calendar Quarter Changes in Psychological Distress The broadly defined lockdown and post-lockdown periods may have smoothed out short-term shifts in psychological distress during the pandemic. To capture these variations, this section re-estimates Eq. ( 1 ) using the preferred specification from Column (2) of Table 3 , detailing changes in mental health by calendar quarter from the first quarter of 2020 (2020Q1) to the fourth quarter of 2021 (2021Q4), when the COVID-19 survey provided an expanded sample. Figure 4 presents these results, illustrating quarterly changes in GHQ-12 scores and the prevalence of psychological distress relative to their counterfactual trends. The GHQ-12 score was .43 points (95% CI: 0.17; 0.69), and cases of psychological distress were 5.4 percentage points (95% CI: 3.2; 7.6) above-trend in 2020Q2, during the first UK lockdown. In relative terms, clinically relevant cases of psychological distress exceeded their counterfactual prediction by 14.7%, while the GHQ-12 rose + 7.3% standard deviations. A slight decline in psychological distress followed before another rise over the winter of 2020/2021, coinciding with renewed lockdown measures. The pandemic effect on GHQ-12 scores peaked in 2021Q1 with 0.9 score points above the counterfactual estimate, while the prevalence of GHQ-12 cases was comparable to its level during the initial pandemic shock. Mental health indicators improved after 2021Q1. By the second quarter of 2021 (2021Q2), reported distress levels were statistically indistinguishable from the counterfactual trend. This suggests that the direct effect of the COVID-19 pandemic on psychological distress was concentrated in the periods of stringent lockdown measures. The full return of psychological distress levels to the predicted counterfactual trend also supports the parallel trends assumption. This finding suggests that, in the absence of the COVID-19 shock, mental health outcomes would have continued along their pre-pandemic trajectory, reinforcing the credibility of the identification strategy. The results confirm that the impact of the pandemic on young adults' mental health was most pronounced in periods of widespread restrictions, with no evidence of persistent mental health deterioration at the population level beyond that point. These results are consistent with the broader trend analysis, which indicates that the rise in psychological distress associated with COVID-19 was temporary and that distress levels returned to their pre-pandemic trajectory once restrictions were permanently lifted, starting in 2021Q2. Heterogeneous Effects The mental health impact of the COVID-19 pandemic may have varied across demographic and socioeconomic groups due to, for example, differences in resources, behaviour, typical activities, or the level of disruptions. However, re-estimating Eq. ( 1 ), including adjustments for survey response effects and time trends in pooled panel samples of 16–59-year-olds, does not suggest notable differences in the mental health response by age over the lockdown and post-lockdown period, as depicted in Fig. 5 . On average across age groups, GHQ-12 was 0.46 points (95% CI [0.36; 0.56]) higher than expected during the acute phase and − 0.11 score points (95% CI [− .21; − .004]) below trend in the recovery phase. Similarly, clinically relevant cases of psychological distress were 4.0 percentage points (95% CI: 3.2; 4.8) above their counterfactual prediction in the acute phase and indistinguishable from trend thereafter (-0.005, 95% CI: − .013; 0.003). Wald tests fail to reject the null hypotheses of age-homogenous mental health responses in response to the lockdown and post-lockdown periods across age groups (GHQ-12 score, p = 0.757; GHQ-12 cases, p = 0.147). Within the sample of young adults, we also tested for differences in COVID-19 effects across gender (male/ female), age (16–21, 22–29), ethnicity (white/ ethnic minorities), baseline household income rank (bottom two-thirds, top third), and economic activity (not in work/ in work). The results indicate some variation in the extent to which different subgroups experienced distress during the lockdown period of the pandemic (see Tables A1 and A2 in the supplement). For GHQ-12 scores, we note a significantly higher psychological distress effect of the lockdown period on young women and respondents in the top third of the household income distribution than the rest. In contrast, young people in work and ethnic minorities experienced a lower effect of the lockdown period on psychological distress than their counterparts (see Table A1 in the supplement); although in the case of ethnic minority groups, the difference did not reach the 5% level of significance. All subgroups’ GHQ-12 sum scores were statistically indistinguishable from their counterfactual trend post-lockdown. Findings for GHQ-12 cases confirm patterns of heterogeneity by gender (women were hit harder) and ethnic minority, but not by economic activity (Table A2 in the supplement). Additional Analyses and Robustness Checks We assessed COVID-19’s effect on life satisfaction and reported loneliness using Eq. ( 1 ), finding minimally higher-than-expected life satisfaction and lower-than-expected loneliness during the post-lockdown period. There is no evidence that either was adversely affected by the COVID-19 pandemic during lockdown (Table A3 in the supplement reports). The estimates confirm the improvements in psychological well-being with receding feelings of loneliness post-lockdown. It is conceivable that the relatively small mental health effects are due to individuals in distress selecting out of the study. Like any panel study, UKHLS suffers from attrition over time. If this is the case, we might underestimate the impact of COVID-19 on mental ill-health. Therefore, we added indicator variables measuring next-wave retention – a respondent participated in the following wave and previous-wave retention – a respondent had participated in the previous wave – in turn to Eq. ( 1 ). These tests were conducted using both continuous psychological distress scores (GHQ-12 score) and clinically relevant cases of distress (GHQ-12 cases). The results for all tests were statistically insignificant (Table 4 ), meaning that panel retention did not predict psychological distress. These results do not suggest that panel retention was systematically related to mental health status. Table 4: Added Variable Test for Panel Retention. GHQ-12 scores GHQ-12 cases Next-wave retention F(1, 14015) = 0.01, p = 0.930 F(1,14015) = 0.01, p = 0.907 Previous-wave retention F(1, 14015) = 0.13, p = 0.721 F(1,14015) = 0.17, p = 0.679 Note: F test results from variables measuring next-wave / previous-wave retention added to Equation (1) Source: UKHLS, UKHLS COVID-19 Study. Author’s calculations. Our identification of COVID-19-related mental health consequences relies on an appropriately fitted time trend, which can be difficult, especially towards the endpoints where data is sparse. To assess the robustness of the findings against an alternative specification, we re-estimated Equation (1) with a restricted cubic spline in the interview date with five knots. Spline functions are piecewise-defined polynomials that are combined in such a way that they are smooth at the points where the pieces join, called knots. They are used in regression analysis to model nonlinear relationships between variables (Perperoglou et al., 2019). A restricted cubic spline is linear before the first knot and after the last knot, which can help extrapolation. The estimates suggest 0.50 points (p < 0.001) higher than expected GHQ-12 score on average during the lockdown phase and a score of -0.14 (p = 0.125) statistically indistinct from trend after that. The prevalence of clinically relevant psychological distress was raised by 4.3 percentage points (p < 0.001) compared to the counterfactual prediction in the lockdown period and statistically indistinct from its trend value afterwards (-.013, p = 0.07), confirming the modest and time-limited effect of the pandemic on young people’s mental health as measured by the GHQ-12 instrument. Finally, to ascertain how well the approach is able to separate out shocks from trends, we conducted a placebo test whereby we ‘switched on’ dummies for the more ‘normal’ years 2017 (Grenfell Tower Fire, Corbyn’s defeat in the General Election, and #MeToo UK) and 2018 (Windrush scandal, royal wedding between Harry and Meghan, and arrival of TikTok), instead of the lockdown and post-lockdown dummy. The results for pseudo-treatments were statistically insignificant at common levels (Table A4), supporting the validity of the parallel trends assumption. Discussion and Conclusion Discussion This study examines COVID-19’s short and long-term mental health impact on 16-29-year-olds in the United Kingdom against longer-term trends, drawing on nationally representative longitudinal data from 2001 to 2023. The data enables a focus on pre-existing trends and experiences during and after the COVID-19 pandemic, controlling for variations in mental health over the life course and potential survey response effects between different survey sources during the pandemic. The findings suggest that already before the COVID-19 pandemic, there had been a rise in mental distress among young people. COVID-19 temporarily accelerated this trend towards mental ill-health, followed by a recovery, with no adverse long-term pandemic-related impacts observed on average, consistent with receding feelings of loneliness and above-trend life satisfaction post-lockdowns. The results apply after adjusting for time-constant individual differences, composition, a non-linear time trend, seasonality, and survey response effects between the UKHLS main and COVID-19 surveys employing linear fixed effects regression models. The study provides much-needed evidence on the longer-term mental health trends among young adults in times of global upheavals. The increased levels of mental distress during the acute phase of the pandemic accelerated pre-existing trends, marked by the 2008 Great Recession, the subsequent UK government austerity programme, and an initial shock reaction in 2020 to the COVID-19 pandemic and lockdown measures, which then subsided after March 2021. Future research has yet to unpack the underlying changes driving the deterioration in young adults' mental health, which may include increases in social media use (Blanchflower, Bryson, et al., 2024 ), declining career outlooks, reduced real income, as well as new living arrangements with parents, partners, and others (Gagné et al., 2022 ), but also cuts to government spending for transport and youth services (Brown et al., 2024 ). Focusing on experiences from 2020 to 2022, the findings support other studies that show that UK young adults experienced historically high levels of distress during the first national lockdown between April and July 2020. However, the findings suggest the initial mental health shock was i) smaller than often reported and ii) temporary after adjusting for pre-existing trends and different response patterns between data sources. The pandemic impact on mental health was stronger among women and young adults in the top third of the household income distribution and less impactful for ethnic minorities, echoing findings by Miall et al. ( 2023 ) for children in the UK. There was no evidence for a stronger pandemic-related mental health decline among young than older adults above 30 years. Similarly, we find no adverse long-term consequences of the pandemic on life satisfaction and reported loneliness on average. The moderate, temporary COVID-19-related setback in mental health is consistent with findings for youths in Norway (Kozák et al., 2023 ) or European adults, more generally (Blanchflower & Bryson, 2024 ). Heterogeneity in the mental health response within young adults during the COVID-19 pandemic might be due to differences in usual social interactions, spare time activities, participation in education and/or employment, or familiarity with stress and uncertainty. However, post-lockdown, average levels of psychological distress were back to their counterfactual trend. The results highlight that, on average, young adults have had the capacity to adapt to the upheaval of the COVID-19 pandemic. However, some adverse COVID-related experiences may continue to predict mental health problems beyond the short and medium term, underscoring the importance of considering individual-specific experiences (Anders & Holt-White, 2024 ). Furthermore, the analysis highlights the magnitude of the mental health burden already present in the UK’s young adult population in the years leading up to 2020. Evidence of resilience during the pandemic suggests that while immediate support and interventions during the lockdown phases of crises were critical, long-term policies should focus on addressing the underlying deterioration in mental health among young adults. Strengthening mental health services and support systems and addressing the root causes to tackle the long-term rise in mental health problems is crucial for improving overall well-being and functioning. Strengths and limitations The study combines nationally representative longitudinal data, including a validated measure of mental health, with a plausible strategy to estimate the causal effects of the COVID-19 pandemic in the short and long term. However, despite its strengths, there are a few possible limitations. First, while the observed trend compares well with patterns in prescriptions for medicines used in mental health, self-reported measures of psychological distress may introduce reporting biases. Second, the estimation of long-term mental health trends relies on a stepwise, coarsened specification of age to remain identifiable. This simplification could introduce bias. Third, while the study adjusts for pre-pandemic trends, life stages, and individual fixed and survey response effects, other unobserved factors could influence mental health outcomes. Fourth, while there is no immediate evidence for violation of the parallel trends assumption that informs the interpretation of the mode parameters as ATT, any unaccounted deviation from the counterfactual trend that happened at the same time as the country entered the lockdown and post-lockdown phase might introduce bias. However, factors such as anticipation effects, seem unlikely and are not borne out in the data (Fig. 4 ). Fifth, the study does not disentangle the direct effect of the pandemic (e.g., mortality and morbidity risks) from indirect effects stemming from containment policies. Finally, attrition across UKHLS has been high among young adults, and the COVID-19 survey waves had relatively low response rates and small young adult samples, introducing selection bias and the threat of underpowered subgroup comparisons. The sensitivity tests, including checks for panel attrition, support the robustness of the findings but cannot entirely eliminate these limitations. Conclusion Despite these limitations, this study offers unique evidence of trends in mental distress among young people in the UK over the past twenty-one years, including before, during and after the COVID-19 pandemic. Rising mental health problems among youth were already observed before the pandemic. We highlighted the increase in distress that young people have faced over time. Going back to trend levels of mental health should, therefore, not suffice as a public health target. The findings emphasise the need for systemic efforts to address the mental health problems among young adults and efforts to promote their well-being in the long term. Relevant initiatives must consider that despite the recovery following the COVID-19 pandemic, pockets of increased vulnerabilities might compound into potential scarring effects on future outcomes. Declarations Data Sharing and Ethical Approval The data used in this study are available from the UK Data Service (study numbers 6614 and 8644). The University of Essex Ethics Committee granted ethical approval for data collection. Syntax replicating analyses, tables and figures is available at https://doi.org/10.5522/04/28469006.v1. Funding There was no funding source for this study. Conflict of interest We declare no competing interests. Acknowledgements: This paper was made possible by data collection funded by the UK Research and Innovation via their COVID response fund and the Economic and Social Research Council. We are grateful for Lulei (Shirley) Chen’s contribution to an earlier draft. References Ahmed, N., Barnett, P., Greenburgh, A., Pemovska, T., Stefanidou, T., Lyons, N., Ikhtabi, S., Talwar, S., Francis, E. R., Harris, S. M., Shah, P., Machin, K., Jeffreys, S., Mitchell, L., Lynch, C., Foye, U., Schlief, M., Appleton, R., Saunders, K. R. K., . . . Johnson, S. (2023). Mental health in Europe during the COVID-19 pandemic: a systematic review. Lancet Psychiatry , 10 (7), 537-556. https://doi.org/10.1016/S2215-0366(23)00113-X Anaya, L., Howley, P., Waqas, M., & Yalonetzky, G. (2023). Locked down in distress: A quasi‐experimental estimation of the mental‐health fallout from the COVID‐19 pandemic. Economic Inquiry , 62 (1), 56-73. https://doi.org/10.1111/ecin.13181 Anders, J., & Holt-White, E. (2024). Young people’s subjective wellbeing in the wake of the COVID-19 pandemic: evidence from a representative cohort study in England. CEPEO Working Paper Series , 24-05 . https://ideas.repec.org/p/ucl/cepeow/24-05.html (UCL Centre for Education Policy & Equalising Opportunities) Anyaegbu, G. (2010). Using the OECD equivalence scale in taxes and benefits analysis. Economic & Labour Market Review , 4 (1), 49-54. https://doi.org/10.1057/elmr.2010.9 Baksheev, G. N., Robinson, J., Cosgrave, E. M., Baker, K., & Yung, A. R. (2011). Validity of the 12-item General Health Questionnaire (GHQ-12) in detecting depressive and anxiety disorders among high school students. Psychiatry Res , 187 (1-2), 291-296. https://doi.org/10.1016/j.psychres.2010.10.010 Banks, J., & Xu, X. (2020). The Mental Health Effects of the First Two Months of Lockdown during the COVID‐19 Pandemic in the UK. Fiscal Studies , 41 (3), 685-708. https://doi.org/10.1111/1475-5890.12239 Blanchflower, D. G. (2025). Declining Youth Well-being in 167 UN Countries. Does Survey Mode, or Question Matter? NBER WORKING PAPER (33415). https://doi.org/10.3386/w33415 Blanchflower, D. G., & Bryson, A. (2024). Were COVID and the Great Recession well-being reducing? PLOS ONE , 19 (11), e0305347. https://doi.org/10.1371/journal.pone.0305347 Blanchflower, D. G., Bryson, A., Lepinteur, A., & Piper, A. (2024). Further Evidence on the Global Decline in the Mental Health of the Young. NBER WORKING PAPER (32500). https://doi.org/10.3386/w32500 Blanchflower, D. G., Bryson, A., & Xu, X. (2024). The Declining Mental Health Of The Young And The Global Disappearance Of The Hump Shape In Age In Unhappiness. NBER WORKING PAPER . https://doi.org/10.3386/w32337 Brown, H., Gao, N., & Song, W. (2024). Regional trends in mental health inequalities in young people aged 16–25 in the UK and the role of cuts to local government expenditure: Repeated cross-sectional analysis using the British household panel Survey/UK household longitudinal survey. Social Science & Medicine , 353 , 117068. https://doi.org/10.1016/j.socscimed.2024.117068 Burton, J., Lynn, P., & Benzeval, M. (2020). How understanding society: the UK household longitudinal study adapted to the COVID-19 pandemic. Survey Research Methods , 14 (2), 235-239. https://doi.org/10.18148/srm/2020.v14i2.7746 Chaudhuri, K., & Howley, P. (2022). The impact of COVID-19 vaccination for mental well-being. European Economic Review , 150 , 104293. https://doi.org/10.1016/j.euroecorev.2022.104293 Conti, G., & Pudney, S. (2011). Survey Design and the Analysis of Satisfaction. The Review of Economics and Statistics , 93 (3). https://doi.org/10.1162/REST_a_00202 Davillas, A., de Oliveira, V. H., & Jones, A. M. (2023). Is inconsistent reporting of self-assessed health persistent and systematic? Evidence from the UKHLS. Economics & Human Biology , 49 , 101219. https://doi.org/10.1016/j.ehb.2022.101219 Dhensa-Kahlon, R. K., Wan, S. T., Coyle-Shapiro, J. A.-M., & Teoh, K. R.-H. (2025). The mental health impact of repeated COVID-19 enforced lockdowns in England: evidence from the UK Household Longitudinal Study. BJPsych Open , 11 (1). https://doi.org/10.1192/bjo.2024.803 Duarte Neves, H., Asaria, M., & Stabile, M. (2024). Young, Muslim and poor: The persistent impacts of the pandemic on mental health in the UK. Social Science & Medicine , 353 , 117032. https://doi.org/10.1016/j.socscimed.2024.117032 Fancourt, D., Steptoe, A., & Bu, F. (2021). Trajectories of anxiety and depressive symptoms during enforced isolation due to COVID-19 in England: a longitudinal observational study. The Lancet Psychiatry , 8 (2). https://doi.org/10.1016/S2215-0366(20)30482-X Fosse, E., & Winship, C. (2019). Analyzing Age-Period-Cohort Data: A Review and Critique. Annual Review of Sociology , 45 (Volume 45, 2019), 467-492. https://doi.org/https://doi.org/10.1146/annurev-soc-073018-022616 Gagné, T., Sacker, A., & Schoon, I. (2022). Transition milestones and life satisfaction at ages 25/26 among cohorts born in 1970 and 1989–90. Advances in Life Course Research , 51 , 100463. https://doi.org/10.1016/j.alcr.2022.100463 Gagné, T., Schoon, I., McMunn, A., Sacker, A., Gagné, T., Schoon, I., McMunn, A., & Sacker, A. (2021). Mental distress among young adults in Great Britain: long-term trends and early changes during the COVID-19 pandemic. Social Psychiatry and Psychiatric Epidemiology 2021 57:6 , 57 (6). https://doi.org/10.1007/s00127-021-02194-7 Henseke, G., Green, F., Schoon, I., Henseke, G., Green, F., & Schoon, I. (2022). Living with COVID-19: Subjective Well-Being in the Second Phase of the Pandemic. Journal of Youth and Adolescence 2022 51:9 , 51 (9). https://doi.org/10.1007/s10964-022-01648-8 Hernán, M. A. (2018). The C-Word: Scientific Euphemisms Do Not Improve Causal Inference From Observational Data. American Journal of Public Health , 108 (5), 616-619. https://doi.org/10.2105/ajph.2018.304337 Hotopf, M., Bullmore, E., O'Connor, R. C., & Holmes, E. A. (2020). The scope of mental health research during the COVID-19 pandemic and its aftermath. The British Journal of Psychiatry , 217 (4), 540-542. https://doi.org/10.1192/bjp.2020.125 Institute for Social and Economic Research. (2021a). Understanding Society COVID-19 User Guide. Version 10.0. In. Colchester: University of Essex. Institute for Social and Economic Research. (2021b). Understanding Society: COVID-19 Study, 2020-2021 (8644; Version 11th Edition). https://doi.org/10.5255/UKDA-SN-8644-11 Institute for Social and Economic Research. (2023). Understanding Society. [data series]. (2000053; Version 11th Release) UK Data Service. https://doi.org/10.5255/UKDA-Series-2000053 Jaschke, P., Kosyakova, Y., Kuche, C., Walther, L., Goßner, L., Jacobsen, J., Ta, T. M. T., Hahn, E., Hans, S., & Bajbouj, M. (2023). Mental health and well-being in the first year of the COVID-19 pandemic among different population subgroups: evidence from representative longitudinal data in Germany. BMJ Open , 13 (6). https://doi.org/10.1136/bmjopen-2022-071331 Kozák, M., Bakken, A., von Soest, T., Kozák, M., Bakken, A., & von Soest, T. (2023). Psychosocial well-being before, during and after the COVID-19 pandemic: a nationwide study of more than half a million Norwegian adolescents. Nature Mental Health 2023 1:7 , 1 (7). https://doi.org/10.1038/s44220-023-00088-y Kung, C. S. J., Kunz, J. S., & Shields, M. A. (2023). COVID-19 lockdowns and changes in loneliness among young people in the U.K. Social Science & Medicine , 320 , 115692. https://doi.org/10.1016/j.socscimed.2023.115692 Lundberg, I., Johnson, R., & Stewart, B. M. (2021). What Is Your Estimand? Defining the Target Quantity Connects Statistical Evidence to Theory. American Sociological Review , 86 (3), 532-565. https://doi.org/10.1177/00031224211004187 Lundin, A., Hallgren, M., Theobald, H., Hellgren, C., & Torgen, M. (2016). Validity of the 12-item version of the General Health Questionnaire in detecting depression in the general population. Public Health , 136 , 66-74. https://doi.org/10.1016/j.puhe.2016.03.005 McManus, S., Gunnell, D., McManus, S., & Gunnell, D. (2019). Trends in mental health, non‐suicidal self‐harm and suicide attempts in 16–24-year old students and non-students in England, 2000–2014. Social Psychiatry and Psychiatric Epidemiology 2019 55:1 , 55 (1). https://doi.org/10.1007/s00127-019-01797-5 Miall, N., Pearce, A., Moore, J. C., Benzeval, M., & Green, M. J. (2023). Inequalities in children’s mental health before and during the COVID-19 pandemic: findings from the UK Household Longitudinal Study. J Epidemiol Community Health , 77 (12), 762-769. https://doi.org/10.1136/jech-2022-220188 Neugebauer, M., Patzina, A., Dietrich, H., & Sandner, M. (2023). Two pandemic years greatly reduced young people’s life satisfaction: evidence from a comparison with pre-COVID-19 panel data. European Sociological Review . https://doi.org/10.1093/esr/jcad077 Niedzwiedz, C. L., Green, M. J., Benzeval, M., Campbell, D., Craig, P., Demou, E., Leyland, A., Pearce, A., Thomson, R., Whitley, E., & Katikireddi, S. V. (2021). Mental health and health behaviours before and during the initial phase of the COVID-19 lockdown: longitudinal analyses of the UK Household Longitudinal Study. J Epidemiol Community Health , 75 (3). https://doi.org/10.1136/jech-2020-215060 O'Connor, R. C., Wetherall, K., Cleare, S., McClelland, H., Melson, A. J., Niedzwiedz, C. L., O'Carroll, R. E., O'Connor, D. B., Platt, S., Scowcroft, E., Watson, B., Zortea, T., Ferguson, E., & Robb, K. A. (2021). Mental health and well-being during the COVID-19 pandemic: longitudinal analyses of adults in the UK COVID-19 Mental Health & Wellbeing study | The British Journal of Psychiatry | Cambridge Core. The British Journal of Psychiatry , 218 (6). https://doi.org/10.1192/bjp.2020.212 Perperoglou, A., Sauerbrei, W., Abrahamowicz, M., & Schmid, M. (2019). A review of spline function procedures in R. BMC Medical Research Methodology , 19 (1), 46. https://doi.org/10.1186/s12874-019-0666-3 Pevalin, D. J. (2000). Multiple applications of the GHQ-12 in a general population sample: an investigation of long-term retest effects. Social Psychiatry and Psychiatric Epidemiology , 35 (11), 508-512. https://doi.org/10.1007/s001270050272 Pierce, M., Hope, H., Ford, T., Hatch, S., Hotopf, M., John, A., Kontopantelis, E., Webb, R., Wessely, S., McManus, S., & Abel, K. M. (2020). Mental health before and during the COVID-19 pandemic: a longitudinal probability sample survey of the UK population. Lancet Psychiatry , 7 (10), 883-892. https://doi.org/10.1016/S2215-0366(20)30308-4 Serrano-Alarcón, M., Kentikelenis, A., Mckee, M., & Stuckler, D. (2022). Impact of COVID‐19 lockdowns on mental health: Evidence from a quasi‐natural experiment in England and Scotland. Health Economics , 31 (2). https://doi.org/10.1002/hec.4453 Settersten, R. A., Jr., Bernardi, L., Harkonen, J., Antonucci, T. C., Dykstra, P. A., Heckhausen, J., Kuh, D., Mayer, K. U., Moen, P., Mortimer, J. T., Mulder, C. H., Smeeding, T. M., van der Lippe, T., Hagestad, G. O., Kohli, M., Levy, R., Schoon, I., & Thomson, E. (2020). Understanding the effects of Covid-19 through a life course lens. Advances in Life Course Research , 45 , 100360. https://doi.org/10.1016/j.alcr.2020.100360 Sun, Y., Wu, Y., Fan, S., Dal Santo, T., Li, L., Jiang, X., Li, K., Wang, Y., Tasleem, A., Krishnan, A., He, C., Bonardi, O., Boruff, J. T., Rice, D. B., Markham, S., Levis, B., Azar, M., Thombs-Vite, I., Neupane, D., . . . Thombs, B. D. (2023). Comparison of mental health symptoms before and during the covid-19 pandemic: evidence from a systematic review and meta-analysis of 134 cohorts. BMJ , 380 , e074224. https://doi.org/10.1136/bmj-2022-074224 Wright, N., Hill, J., Sharp, H., Refberg-Brown, M., Crook, D., Kehl, S., Pickles, A., Wright, N., Hill, J., Sharp, H., Refberg-Brown, M., Crook, D., Kehl, S., & Pickles, A. (2024). COVID-19 pandemic impact on adolescent mental health: a reassessment accounting for development. European Child & Adolescent Psychiatry . https://doi.org/10.1007/s00787-023-02337-y Zhang, A., Gagne, T., Walsh, D., Ciancio, A., Proto, E., & McCartney, G. (2023). Trends in psychological distress in Great Britain, 1991-2019: evidence from three representative surveys. J Epidemiol Community Health , 77 (7), 468-473. https://doi.org/10.1136/jech-2022-219660 Footnotes There were 167 interviews with the target age-group in 2024, which were assigned to their sampling year-month. Additional Declarations The authors declare no competing interests. Supplementary Files Supplement.docx Cite Share Download PDF Status: Published Journal Publication published 09 May, 2025 Read the published version in Social Indicators Research → Version 1 posted 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-6154489","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":423973958,"identity":"0e99920b-053a-44dd-9e3c-a3216bd3b38e","order_by":0,"name":"Golo Henseke","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA10lEQVRIiWNgGAWjYFACHjBpwMDewMDwAMxmI1YLzwEGhgTStEgkEKlF3oH34OOCPzbG/JJvDD8kMNjJM0ikJeDVYniAL9l4ZluameTsHGOgRcmGDRJpB/BraeAxk+ZtOGxjcDvHDOgw5gQGifQGQlrMf/P8+W9jf/MMSEs9YS3yDDxmzDxsB8wMJHhAWg4DtRBwmAEzj7E0b1uyscSZtGKJBIPjhm08zxLw29LeY/iZ54+dYX/74Y0fPlRUy/Ozpxngt+UwKpeIiJRvIKRiFIyCUTAKRgEADIs5TOFsiDEAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0003-0669-2100","institution":"University College London","correspondingAuthor":true,"prefix":"","firstName":"Golo","middleName":"","lastName":"Henseke","suffix":""},{"id":423974506,"identity":"4a050e85-9d53-4bbe-ac5c-12f762aff136","order_by":1,"name":"Ingrid Schoon","email":"","orcid":"https://orcid.org/0000-0002-4262-3711","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Ingrid","middleName":"","lastName":"Schoon","suffix":""}],"badges":[],"createdAt":"2025-03-04 12:30:39","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6154489/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6154489/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11205-025-03616-8","type":"published","date":"2025-05-10T00:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":77761799,"identity":"fe083f90-01ae-4e73-9188-87dcd36d6879","added_by":"auto","created_at":"2025-03-05 09:27:21","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":613508,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of UKHLS Samples over Calendar Quarters, 2020-2023\u003c/p\u003e\n\u003cp\u003eNote: Count of cases in the UKHLS COVID-19 sample by data source over 2020-2023. Author’s calculations.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6154489/v1/ee5e9d2d52d9a942cea06292.png"},{"id":77761800,"identity":"119e1049-3593-4ce3-83d0-86f54644896a","added_by":"auto","created_at":"2025-03-05 09:27:21","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":548940,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of the GHQ-12 score in the Covid-19 Sample 2009-2023\u003c/p\u003e\n\u003cp\u003eNote: Pooled frequency distribution of the GHQ-12 scores in BHPS, UKHLS, and UKHLS Covid-19 study. Author’s calculations.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6154489/v1/324a8c39bdae3643c86c9548.png"},{"id":77761806,"identity":"e12c36c5-0e73-4e4f-81db-c4c1d260db6c","added_by":"auto","created_at":"2025-03-05 09:27:22","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":795962,"visible":true,"origin":"","legend":"\u003cp\u003eLong-term Changes in Psychological Distress, 2001-2023\u003c/p\u003e\n\u003cp\u003eNote: Time trend in GHQ-12 scores and cases of psychological distress estimated using Equation (1), incorporating individual fixed effects, year dummies, and age group controls. Dummy variables for the COVID-19 study interacted with the 2021-year dummy are included. The shaded areas represent 95% CI. \u0026nbsp;N = 120,987 (n = 22,247).\u003c/p\u003e\n\u003cp\u003eSources: BHPS, UKHLS Mainstage, and COVID-19 study. Authors’ calculations.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6154489/v1/b9b5dafd06d98ae3ef2711fe.png"},{"id":77761803,"identity":"b7439b1d-11bf-4f08-8a43-65b3314d2275","added_by":"auto","created_at":"2025-03-05 09:27:21","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":603760,"visible":true,"origin":"","legend":"\u003cp\u003eThe Evolution of Young Adults’ Mental Health by Calendar Quarter, 2020-2021\u003c/p\u003e\n\u003cp\u003eNote: Charts depict the average marginal effect of quarterly dummies from linear fixed-effects regressions of Equation (1) of the GHQ-12 score and psychological distress indicator with adjustments for pre-existing trends, life course, seasonal variations, and survey dummies interacted with interview year-quarter dummies in the COVID-19 sample of 16–29-year-olds. N=96,564 (n=18,027). 95% CI included.\u003c/p\u003e\n\u003cp\u003eSource: UKHLS, UKHLS COVID-19 survey. Authors’ calculations.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6154489/v1/52797ae150047be9fbc4e3c2.png"},{"id":77760349,"identity":"a29042d4-83b5-4674-91bd-725c2e9ab1b5","added_by":"auto","created_at":"2025-03-05 09:19:21","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":493523,"visible":true,"origin":"","legend":"\u003cp\u003eLockdown and Post-Lockdown Effects on the GHQ-12 Score and GHQ-12 Cases by Age.\u003c/p\u003e\n\u003cp\u003eNote: Charts depict the average marginal effect for the lockdown and post-lockdown dummy variables from linear fixed-effects regressions of Equation (1) of the GHQ-12 score and GHQ-12 cases indicator with adjustments for pre-existing trends, life course, time trend, seasonal variations, and a survey dummy interacted with the interview year-quarter dummies in samples of 16–59-year-olds (N = 397,349). Fully interacted with age-group indicators. The figure includes 95% CI.\u003c/p\u003e\n\u003cp\u003eSource: UKHLS, UKHLS COVID-19 survey. Authors’ calculations.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-6154489/v1/b48336eee94f755234337431.png"},{"id":82551061,"identity":"956078ba-d3e8-4690-88c9-54c310b7fe1a","added_by":"auto","created_at":"2025-05-12 20:15:04","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3829300,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6154489/v1/c8da4322-48d1-4d01-9395-ea25710ff8df.pdf"},{"id":77760347,"identity":"24245a12-e976-4a44-a377-e8c484575e25","added_by":"auto","created_at":"2025-03-05 09:19:21","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":23432,"visible":true,"origin":"","legend":"","description":"","filename":"Supplement.docx","url":"https://assets-eu.researchsquare.com/files/rs-6154489/v1/f848d7e1813ff07909ec1aaa.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eRevisiting the Mental Health Impact of COVID-19 on Young Adults in the UK: Long-Term Trends, Temporary Setbacks, and Recovery\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThis paper assesses the mental health of 16–29-year-olds in the United Kingdom between 2001 and 2023 before, during, and after the COVID-19 pandemic, focusing on the \u003cem\u003ecausal effects\u003c/em\u003e of the lockdown and post-lockdown period (Hernán, \u003cspan class=\"CitationRef\"\u003e2018\u003c/span\u003e). It analyses long-term trends in psychological distress and evaluates the pandemic’s impact during the lockdown and post-lockdown period, accounting for differences in self-reported mental health across UK Household Longitudinal Study (UKHLS) surveys.\u003c/p\u003e\n\u003cp\u003eFrom the outset, there were major concerns about the far-reaching mental health effects of COVID-19 (Hotopf et al., \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e). For example, Ahmed et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) counted 177 studies on pandemic-related mental health outcomes in Europe alone. While young people were less vulnerable to the disease’s direct health risks, stringent containment measures disrupted their education, careers, leisure activities, and peer relationships, potentially affecting their developmental trajectories (Settersten et al., \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e). Numerous studies reported a pronounced deterioration in mental health during the pandemic’s early phase in the UK (Fancourt et al., \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e), with young people and women disproportionately affected (Banks \u0026amp; Xu, \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e). However, research had to rely on convenience samples (O'Connor et al., \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e), did not adjust for pre-existing trends in mental health (Niedzwiedz et al., \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e), or did not consider how differences between surveys might have influenced self-reported mental health (Anaya et al., \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e; Pierce et al., \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e), potentially biasing estimates of pandemic-related impacts.\u003c/p\u003e\n\u003cp\u003eIndeed, longitudinal evidence outside the UK represents a less clear-cut picture of elevated mental health risks (e.g., Jaschke et al., \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). More recent global reviews of high-quality longitudinal evidence suggest that mental health remained unchanged or worsened only minimally during the pandemic—even among adolescents and young adults (Sun et al., \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). These findings contrast with narratives of a widespread and lasting decline in population well-being. Moreover, measurement issues have received relatively little attention, even though many UK-based studies rely on pre- and post-pandemic comparisons using different surveys.\u003c/p\u003e\n\u003cp\u003eFurthermore, evidence on mental health recovery following the vaccine rollouts, the lifting of lockdown restrictions, and economic rebounds remain limited. While receding stressors might be expected to improve young adults’ psychological well-being, prolonged and repeated disruptions could have developed lasting ‘scarring’ effects on psychosocial development. Existing research presents mixed findings, reflecting context-specific mental health trajectories and differences in how well-being is conceptualised. For example, in the UK, Gagné et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) documented a sharp increase in the risk of psychological distress (GHQ-12) between 2018/19 and April 2020, followed by a decline between July and September 2020 in individuals aged 16–34. Similarly, Henseke et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) observed a substantial rise in life satisfaction between February 2021 and May 2022, linked to increased social interactions and reduced uncertainties about learning and careers. The vaccine rollout also improved mental health, particularly among clinically vulnerable groups (Chaudhuri \u0026amp; Howley, \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). Feelings of loneliness had returned to pre-pandemic levels by September 2021 (Kung et al., \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). In Norway, Kozák et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) report above-average scores of depressive symptoms among adolescents in 2021 but not in 2022, suggesting a recovery of mental health to long-term trends. In contrast, Neugebauer et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) reported persistent declines in life satisfaction in a sample of German high school students, and Dhensa-Kahlon et al. (\u003cspan class=\"CitationRef\"\u003e2025\u003c/span\u003e) found elevated psychological distress in UK adults aged 18–29 post-lockdown, though the extent to which these changes were pandemic-driven remains unclear.\u003c/p\u003e\n\u003cp\u003eThis contrasting evidence underscores the importance of contextualising pandemic-related mental health trends within pre-existing dynamics. A growing body of research indicates that young people’s mental health has been declining since at least the 2008 Great Recession (Blanchflower, Bryson, et al., \u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e), as reflected in lower life satisfaction (Gagné et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e), more reported symptoms of distress (Zhang et al., \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e), and rising self-harm (McManus et al., \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e). Irrespective of the causes, this ongoing deterioration suggests that distress levels in 2020 and beyond may have changed due to pre-pandemic trends rather than the pandemic itself. Additionally, psychological well-being tends to decline during the transition to adulthood (Blanchflower, Bryson, \u0026amp; Xu, \u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e), explaining part of the mental health trends observed during the pandemic (Wright et al., \u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e). Finally, small changes in survey design and administration can influence well-being estimates (Blanchflower, \u003cspan class=\"CitationRef\"\u003e2025\u003c/span\u003e; Conti \u0026amp; Pudney, \u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e; Davillas et al., \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e), complicating longitudinal comparisons across studies.\u003c/p\u003e\n\u003cp\u003eUsing longitudinal survey data from the UK Household Longitudinal Study (UKHLS, also known as Understanding Society), this study makes several key contributions. Firstly, we compare self-reported mental health between the UKHLS main survey and the COVID-19 study to evaluate whether questionnaire design influenced reported distress levels. Secondly, we track changes in psychological distress among 16–29-year-olds since 2001, adjusting for individual fixed effects (including cohort effects) and age-related life course changes (e.g., employment and household composition). Thirdly, we use a Neyman-Rubin causal approach to identify the average treatment effect on the treated of the COVID-19 pandemic, estimating individual-specific counterfactual levels of distress had 2020 and 2021 been ‘normal’ years, while accounting for differences between surveys.\u003c/p\u003e\n\u003cp\u003eThis study aligns with prior UKHLS-based research on pandemic-related mental health impacts. Pierce et al. (\u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e) seminal contribution assessed the change in mental health in April 2020 against a counterfactual prediction of what would have been expected from population trends up to 2018/2019, finding a sharp increase in GHQ-12 distress among individuals under 35. However, their estimates did not account for age-specific mental health trends. Banks and Xu (\u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e) adjusted for these trends, confirming disproportionate mental health impacts on young adults and women. Gagné et al. (\u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e) showed that distress declined unevenly after April 2020, with slower recovery among women and young adults aged 25–34. More recently, Anaya et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) used a difference-in-differences approach, finding a + 32% standard deviation increase in GHQ-12 scores among 18–34-year-olds during the first national lockdown. Serrano-Alarcón et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) identified a reduction in mental ill-health from the easing of containment measures, pointing towards greater responsiveness of mental health to the behavioural restrictions than to the virus itself. Duarte Neves et al. (\u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e) also deployed a difference-in-difference approach, finding evidence for long-term ‘scarring’ of mental health from the pandemic for some population groups. All these studies combined data collected through a shortened web questionnaire launched in April 2020 with the UKHLS main survey, assuming rather than assessing measurement equivalence in self-reported mental health problems between the surveys.\u003c/p\u003e\n\u003cp\u003eThe remainder of the paper is structured as follows. The next section introduces the datasets, measures, and the analytical strategy. Section 3 presents and discusses the findings. The final section concludes.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eDatasets\u003c/p\u003e\u003cp\u003eThe analysis draws on combined survey data from the first 14 waves of the UK Household Longitudinal Study (UKHLS), waves 11–18 of its predecessor, the British Household Panel Study (BHPS), and all sweeps of the UKHLS COVID-19 web survey (2020–2021). All data files are available for research at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5255/UKDA-Series-2000053\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e\u003cp\u003eUKHLS is an ongoing panel survey of about 40,000 households in the United Kingdom, launched in 2009 as a continuation of the BHPS, which ran from 1991 to 2008 (Institute for Social and Economic Research, \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). The most recent wave (Wave 14) includes 35,500 individuals, nearly 6,000 of whom are under 30 years old. Fieldwork for each wave spans three years, with samples continually issued during the first two years. For example, UKHLS Wave 1 ran from 2009 to 2011 and Wave 14 issued samples from 2021 to 2023\u003csup\u003e1\u003c/sup\u003e, creating overlapping fieldwork periods in each calendar year. Adults aged 16 and over in sampled households are re-interviewed annually, including core members (initial sample members and their descendants) who move or form a new household.\u003c/p\u003e\u003cp\u003eThe BHPS followed a similar panel study protocol but with notable differences. It was smaller in scale, with full UK-wide coverage starting in 2001/2002 (Wave 11). Fieldwork typically began in September and lasted until April the following year. The final BHPS wave (2008/2009) collected data from 14,400 individuals (3,200 under 30 years old) across 8,100 households. UKHLS continued the BHPS sample from Wave 2 onward.\u003c/p\u003e\u003cp\u003eDue to lockdown restrictions, the UKHLS main survey shifted to a web-first mode in mid-March 2020, with most responses collected online and a smaller proportion via telephone. Face-to-face fieldwork commenced in April 2022. However, even before the pandemic, 70% of panel members were already invited to complete the questionnaire web-first (Burton et al., \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e). This means that the transition away from in-person interviews in 2020 was an acceleration of an existing trend rather than an abrupt shift.\u003c/p\u003e\u003cp\u003eStarting in April 2020, UKHLS participants were invited to complete a 20-minute web survey on their experiences and reactions to the COVID-19 pandemic (Institute for Social and Economic Research, \u003cspan class=\"CitationRef\"\u003e2021b\u003c/span\u003e). The first four waves of the COVID-19 study were conducted monthly between April and July 2020, followed by bi-monthly surveys from September 2020 to March 2021 (waves 5–8). A final ninth wave was conducted in September 2021.\u003c/p\u003e\u003cp\u003eAs with the main survey over this period, responses were collected predominantly online, with a minor telephone mode available in waves 2 and 6. In the under-30 age group, telephone responses accounted for 39 cases in total (0.3% of all COVID-19 study cases under 30), which we integrated into the web survey dataset and did not treat separately. Response rates to the COVID-19 study varied from 42% in wave 1 to 29% in waves 6 and 7 (Institute for Social and Economic Research, \u003cspan class=\"CitationRef\"\u003e2021a\u003c/span\u003e). The COVID-19 study is fully integrated with UKHLS, using shared panel IDs and including core demographic, life course, and mental health information.\u003c/p\u003e\u003cp\u003eWe extracted mental health, demographics, and life course data for 16–29-year-olds across UKHLS and BHPS survey waves since 2001. After removing singleton observations, there were 120,987 person-year observations from 22,247 individuals to measure long-term changes in mental health from 2001–2022 in the age-group 16–29 years.\u003c/p\u003e\u003cp\u003eTo assess the mental health effects of COVID-19, we restrict the sample to cases aged 16–29 years in the UKHLS waves 1–14 (2009–2023) and the COVID-19 study (2020–2021). We employ a complete-case analysis in the unbalanced panel, acknowledging that sample attrition and excluding respondents with missing data may introduce bias. The limitations of this approach and robustness checks assessing attrition bias will be discussed in subsequent sections, where we introduce an added-variable test to examine whether panel retention is systematically associated with mental health outcomes in our estimation model. In all, the COVID-19 sample consists of 96,564 person-wave observations from 18,027 individuals aged 16–29, with an average panel retention of 5.4 waves per respondent. Figure \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e shows the distribution of observations over calendar quarters (2020–2023) and their source. The COVID-19 web surveys contributed the bulk of cases over the pandemic, at 64% in 2020 and 43% in 2021, while UKHLS main survey data provided a consistent reference point for benchmarking changes in reported mental ill-health. To compare patterns with older age groups, some analyses will lift the age restriction to include cases up to age 59 years.\u003c/p\u003e\u003cp\u003eMeasures\u003c/p\u003e\u003cp\u003eMental health is assessed using the 12-item General Health Questionnaire (GHQ-12), a widely used, validated, and reliable instrument for measuring non-specific psychological distress in longitudinal samples (Lundin et al., \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e) and youth populations (Baksheev et al., \u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e). The GHQ-12 has been administered in every wave of the UKHLS, its COVID-19 study and BHPS, ensuring consistent measurement over time. Moreover, prior research has found no evidence of panel conditioning effects in GHQ-12 responses (Pevalin, \u003cspan class=\"CitationRef\"\u003e2000\u003c/span\u003e), further supporting the instrument’s consistency over repeated administrations.\u003c/p\u003e\u003cp\u003eFor the analysis, two complementary measures of psychological distress are used, both derived from the same GHQ-12 instrument. The first is a continuous GHQ-12 score, constructed by summing responses across the twelve items, where each item is scored on a four-point Likert scale: 0 for “not at all,” 1 for “no more than usual,” 2 for “rather more than usual,” and 3 for “much more than usual.” The total GHQ-12 score ranges from 0 to 36, with higher values indicating greater psychological distress. The second measure is a binary indicator of psychological distress derived from the GHQ-12 instrument, identifying individuals experiencing clinically relevant distress if they reported “rather more than usual” or “much more than usual” on at least four of the twelve items. This cutpoint-based classification, commonly used in epidemiological and social science research (Pierce et al., \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e), helps to distinguish clinically relevant cases of psychological distress. Figure\u0026nbsp;2 shows the distribution of the GHQ-12 sum score in the trend sample, demonstrating that the measure effectively captures a broad range of distress levels. The distribution is right-skewed, with most observations concentrated in the lower-to-mid range but with a clear spread toward higher distress levels. Importantly, there is no apparent clustering near the upper end of the scale, confirming that the measure retains sensitivity at higher distress levels. This pattern suggests that ceiling effects are unlikely to be a major concern, as the GHQ-12 scale allows room for capturing worsening psychological distress if present.\u003c/p\u003e\u003cp\u003eIn addition to mental health measures, for subgroup analyses and to control for life course milestones, we extract information on individual age, sex, ethnic minority status, employment status, partnership status, and the number of young children under five in the household. Socioeconomic status is measured using the percentile household income rank at the time individuals entered the panel, based on a comparison of gross monthly household income equivalised for household composition using the modified OECD scale, ranked within each survey wave (Anyaegbu, \u003cspan class=\"CitationRef\"\u003e2010\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTo measure time trends and survey effects, interview date information is incorporated to adjust for long-term changes and seasonal fluctuations in mental health outcomes. A binary indicator for the COVID-19 web study differentiates between the main survey and COVID-19 study observations.\u003c/p\u003e\u003cp\u003eTable \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e presents summary statistics for both the long-term trend sample (2001–2023) and the UKHLS sample (2009–2023). The two samples are highly comparable in terms of demographic composition and key characteristics. The mean GHQ-12 score was 11.4 in the trend sample and 11.6 in the UKHLS sample, with a similar proportion classified as psychologically distressed (23% vs. 24%). Both samples were skewed towards women (57% in the trend sample and 58% in the UKHLS sample), and about one-quarter of respondents identified as belonging to an ethnic minority group (24% vs. 27%). The distribution of age groups was stable across samples, with approximately 30% aged 16–19, 35% aged 20–24, and 35% aged 25–29. Socioeconomic factors also showed minimal differences; the mean household income rank was 49.41 in the trend sample and 48.20 in the UKHLS sample. Similarly, employment rates (54% vs. 53%) and partnership status (27% vs. 26%) were nearly identical.\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eSummary Statistics for Samples of 16-29-year-olds\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\u003cth align=\"left\"\u003e\n \u003cp\u003eN\u003c/p\u003e\n \u003c/th\u003e\u003cth align=\"left\"\u003e\n \u003cp\u003eIndividuals\u003c/p\u003e\n \u003c/th\u003e\u003cth align=\"left\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/th\u003e\u003cth align=\"left\"\u003e\n \u003cp\u003eStd. Dev\u003c/p\u003e\n \u003c/th\u003e\u003cth align=\"left\"\u003e\n \u003cp\u003eStd. Dev (within)\u003c/p\u003e\n \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"6\"\u003e\n \u003cp\u003e\u003cem\u003eLong-Term Trend Sample 2001–2023\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eGHQ-12 (Score)\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e120987\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e22247\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e11.41\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e5.95\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e4.17\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eGHQ-12 (Cases)\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e120987\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e22247\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.23\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.42\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.32\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e120984\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e22246\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.57\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.50\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.00\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eAge: 16–19\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e120987\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e22247\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.30\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.33\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eAge: 20–24\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e120987\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e22247\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.40\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eAge: 25–29\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e120987\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e22247\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.33\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eEthnic minority\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e119012\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e21777\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.24\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.42\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.00\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eHousehold income rank\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e119632\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e21973\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e49.41\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e27.84\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.00\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eIn work\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e120987\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e22247\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.54\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.50\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eLiving as a couple\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e120987\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e22247\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.45\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.24\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eNumber of children \u0026lt; 5 in the household\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e120987\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e22247\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.18\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"6\"\u003e\n \u003cp\u003e\u003cem\u003eUKHLS Sample 2009–2023\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eGHQ-12 (Score)\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e96564\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e18027\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e11.64\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e6.03\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e4.20\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eGHQ-12 (Cases)\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e96564\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e18027\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.24\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.42\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.33\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e96561\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e18026\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.58\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.49\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.00\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eAge: 16–19\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e96564\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e18027\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.30\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.32\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eAge: 20–24\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e96564\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e18027\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.40\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eAge: 25–29\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e96564\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e18027\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.32\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eEthnic minority\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e95044\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e17659\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.44\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.00\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eHousehold income rank\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e95222\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e17757\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e48.20\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e27.68\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.00\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eIn work\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e96564\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e18027\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.53\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.50\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eLiving as a couple\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e96564\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e18027\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.44\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.23\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003eNumber of children \u0026lt; 5 in the household\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e96564\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e18027\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.17\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.45\u003c/p\u003e\n \u003c/td\u003e\u003ctd align=\"left\"\u003e\n \u003cp\u003e0.25\u003c/p\u003e\n \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003eSource: UKHLS and BHPS main surveys, UKHLS COVID-19 web surveys. Sample of 16-29-year-olds. Authors’ calculations.\u003c/p\u003e\u003cp\u003eAnalytical Approach\u003c/p\u003e\u003cp\u003eOur analytical approach estimates the causal impact of the COVID-19 pandemic on young adults’ mental health by comparing observed outcomes during the pandemic with counterfactual outcomes predicted from pre-pandemic trends. We begin with a general model of individual mental health over time:\u003c/p\u003e\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e$$\\:M{H}_{it}={\\alpha\\:}_{it}+{\\gamma\\:}_{t}\\:COVID\\left(t\\right)+{\\gamma\\:}_{s}\\:COVID\\left(s\\right)+{ϵ}_{it}$$\u003c/div\u003e\n\u003c/div\u003e\u003cp\u003eWhere:\u003c/p\u003e\u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cspan class=\"mathinline\"\u003e\\(\\:M{H}_{it}\\)\u003c/span\u003e\u0026nbsp;\u003c/span\u003e represents the mental health outcome (GHQ-12 scores or cases indicator) for individual \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:i\\:\\)\u003c/span\u003e\u003c/span\u003eat time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\)\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cspan class=\"mathinline\"\u003e\\(\\:{\\alpha\\:}_{it}\\)\u003c/span\u003e\u0026nbsp;\u003c/span\u003e is an individual-specific, time-dependent component that reflects life-course variation in predisposition towards mental (ill-)health.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cspan class=\"mathinline\"\u003e\\(\\:{\\gamma\\:}_{t,s}\\:\\)\u003c/span\u003e\u0026nbsp;\u003c/span\u003e captures the effect of the COVID-19 pandemic on mental health, distinguishing between the lockdown period (t = 1) and post-lockdown period (s = 1).\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cspan class=\"mathinline\"\u003e\\(\\:{ϵ}_{it}\\)\u003c/span\u003e\u0026nbsp;\u003c/span\u003e represents idiosyncratic mental health shocks.\u003c/p\u003e\n \u003c/li\u003e\n\u003c/ul\u003e\u003cp\u003eOur ‘estimand’ or target parameter is the Average Treatment Effect on the Treated (ATT) for the lockdown and post-lockdown periods (Lundberg et al., \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e). Specifically, for the lockdown period we define\u003c/p\u003e\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e$$\\:AT{T}_{t}=E\\left[M{H}_{it}^{1}-{MH}_{it}^{0}|COVID\\left(t\\right)=1\\right]$$\u003c/div\u003e\n\u003c/div\u003e\u003cp\u003ewith an analogous definition for the post-lockdown period. This means the pandemic effect is conceptualised as the difference between the observed average mental health outcomes and the counterfactual scenario where the pandemic did not occur.\u003c/p\u003e\u003cp\u003eTo estimate the ATT, we employ a linear fixed effects regression model that accounts for individual-specific time-invariant heterogeneity and flexible time trends. The model is specified as:\u003c/p\u003e\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e$$\\:M{H}_{it}={\\alpha\\:}_{i}+g\\left(t\\right)+{X}_{it}\\beta\\:+{{\\delta\\:}_{t}D}_{it}+{\\gamma\\:}_{t}\\:COVID\\left(t\\right)+{\\gamma\\:}_{s}\\:COVID\\left(s\\right)+{\\epsilon\\:}_{it}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003cp\u003eWhere:\u003c/p\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cspan class=\"mathinline\"\u003e\\(\\:{\\alpha\\:}_{i}\\)\u003c/span\u003e\u0026nbsp;\u003c/span\u003e captures individual-specific fixed effects.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eg(t) is a function of time that captures period trends in mental health. Initially, we model g(t) as a cubic polynomial in survey years to capture long-term trends since 2009, following prior research (Banks \u0026amp; Xu, \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e; Pierce et al., \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e). We later test the robustness by replacing the cubic polynomial with a restricted cubic spline in interview dates.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cspan class=\"mathinline\"\u003e\\(\\:{X}_{it}\\)\u003c/span\u003e\u0026nbsp;\u003c/span\u003e is a vector of time-varying covariates, including age bands (\u0026lt; 20, 20–24, 25+), partnership status, the number of children under 5 in the household, and survey month. These covariates serve as proxies for life-cycle markers, helping to approximate age-related effects and adjust for seasonal fluctuations in mental health.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cspan class=\"mathinline\"\u003e\\(\\:{D}_{it}\\:\\)\u003c/span\u003e\u0026nbsp;\u003c/span\u003etime-varying survey design effects (dummy variable for COVID-19 surveys vs UKHLS main survey), accounting for potential level differences in self-reported mental health across surveys.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cspan class=\"mathinline\"\u003e\\(\\:COVID(t,s)\\)\u003c/span\u003e\u0026nbsp;\u003c/span\u003e are indicator variables for the lockdown and post-lockdown periods, with the corresponding parameter \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\gamma\\:}_{t,s}\\)\u003c/span\u003e\u003c/span\u003e capturing the pandemic’s effects on mental health.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cspan class=\"mathinline\"\u003e\\(\\:{\\epsilon\\:}_{it}\\)\u003c/span\u003e\u0026nbsp;\u003c/span\u003e is the error term, assumed to be independent of the covariates, time trend and individual-fixed effects.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003cp\u003eUnder the assumption that our empirical model is correctly specified—meaning that individual fixed effects, time trends, and covariates fully capture the counterfactual mental health trajectory in the absence of the pandemic—the coefficients \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\gamma\\:}_{t,s}\\)\u003c/span\u003e\u003c/span\u003e in Eq. (\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e) represent the causal impact of the COVID-19 pandemic on mental health. Formally, we have\u003c/p\u003e\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e$$\\:AT{T}_{t,s}=E\\left[M{H}_{i}^{1}\\left(t,s\\right)-{MH}_{i}^{0}\\left(t,s\\right)\\right|COVID(t,s)=1]={\\gamma\\:}_{t,s}\\:$$\u003c/div\u003e\u003c/div\u003e\u003cp\u003ewhich rests on the assumption that the counterfactual mental health —what would have occurred absent the pandemic—is fully captured by the pre-pandemic trends, individual fixed effects, and covariates: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\alpha\\:}_{i}+g\\left(t\\right)+{\\varvec{X}}_{it}\\beta\\:+{{\\delta\\:}_{t}D}_{it}\\)\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e\u003cp\u003eThis approach relies on the \u003cem\u003eparallel trends assumption\u003c/em\u003e, which posits that, in the absence of COVID-19, mental health outcomes during the pandemic would have followed the same underlying trajectory as in the pre-pandemic period. By incorporating individual fixed effects and a flexible time trend, our model effectively imposes this parallel trend assumption. Therefore, if this assumption holds, then any estimated deviations from the expected trajectory during the pandemic—captured by the COVID-19 period indicators—reflect the ATT.\u003c/p\u003e\u003cp\u003eTo strengthen confidence in the parallel trends assumption and address potential threats to identification, we conduct several supplementary analyses and robustness checks:\u003c/p\u003e\u003cp\u003e1. \u003cem\u003eCovariate Adjustment\u003c/em\u003e: While the model assumes that COVID-19’s effects on mental health do not operate through life course markers, this assumption can be contested. We assess sensitivity by estimating the model without these covariates and comparing results.\u003c/p\u003e\u003cp\u003e2. \u003cem\u003eEvent Study\u003c/em\u003e: To examine the evolution of mental health changes over shorter intervals, we break the lockdown period into calendar quarters, allowing us to verify whether pre-pandemic trends closely mirror the dynamics observed during the pandemic in the absence of treatment.\u003c/p\u003e\u003cp\u003e3. \u003cem\u003eSubgroup Analysis\u003c/em\u003e: We conducted subgroup analyses to examine whether the pandemic’s effects varied across different demographic and socioeconomic groups. This helps illuminate the stability of our estimated effect, adding substantive insights as well.\u003c/p\u003e\u003cp\u003e4. \u003cem\u003eAlternative Measures of Subjective Well-Being\u003c/em\u003e: We re-estimate models using alternative measures of subjective well-being (single-item life satisfaction and feelings of loneliness), testing whether findings are robust to different operationalisations of mental health and, thus, potentially different measurement errors.\u003c/p\u003e\u003cp\u003e5. \u003cem\u003eAttrition Bias Testing\u003c/em\u003e: Given that panel dropout could be related to mental health, we explicitly test for attrition bias to ensure that selection effects do not confound estimates.\u003c/p\u003e\u003cp\u003e6. \u003cem\u003eAlternative Time Trend\u003c/em\u003e: We replaced the cubic polynomial in survey years with a restricted cubic spline in interview dates to assess robustness against different specifications of the long-term time trend.\u003c/p\u003e\u003cp\u003e7. \u003cem\u003ePlacebo Test\u003c/em\u003e: Finally, we conducted a placebo test, assigning pseudo-treatment periods to years before the pandemic. The absence of significant effects in these placebo tests reinforces the credibility of our identification strategy.\u003c/p\u003e\u003cp\u003eTogether, these checks address potential threats to identification and provide a framework for interpreting our estimates of the lockdown and post-lockdown effects as the causal impact of the COVID-19 pandemic and associated restrictions on mental health. Additionally, these analyses offer substantive insights into how young people coped with the pandemic and its immediate aftermath. While further discussion of life-cycle (age, period, and cohort) effects and the role of specific covariates is provided in the findings section, it is worth noting here that controlling for partnership status, the presence of young children, and broad age groups along with survey month adjustments, helps to separate period-specific effects from broader life-course dynamics.\u003c/p\u003e\u003cp\u003eAll data cleaning, management, and analyses were conducted in Stata 18.5. Syntax files to replicate analysis, tables and figures are hosted at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5522/04/28469006.v1\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e"},{"header":"Findings","content":"\u003cp\u003eDifferences in Psychological Distress Between the UKHLS Main Survey and the COVID-19 Study\u003c/p\u003e\u003cp\u003ePooling observations from the UKHLS main survey and the COVID-19 study assumes that mental health measures are comparable across surveys. This assumption is based on the use of the same GHQ-12 instrument and the fact that both surveys were administered primarily via web questionnaires during the pandemic. However, if the questionnaire design introduced systematic measurement differences, pooling the two data sources without adjustment may introduce bias in longitudinal analyses of mental health before, during, and after the pandemic. To test for potential discrepancies, this section compares average levels of the GHQ-12 sum score and the prevalence of clinically relevant cases of psychological distress between the main survey and the COVID-19 study conducted in the same survey year-month.\u003c/p\u003e\u003cp\u003eTable \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e presents the results, comparing average differences in reported psychological distress between surveys in the age group 16–29 with figures for 30–44 and 45-59-year-olds. The findings suggest statistically significant differences in reported mental health levels between the two surveys, particularly among younger adults. In the 16–29 age group, the COVID-19 study recorded higher distress levels than the main survey, with an average GHQ-12 score difference of 0.8 points (95% CI [0.46, 1.14]) and a 2.1 percentage point increase in the prevalence of psychological distress (95% CI [-0.002, 0.044]), although the latter was not statistically significant at conventional levels.\u003c/p\u003e\u003cp\u003eFor individuals aged 30–44, the COVID-19 study also reported higher distress levels, with an average GHQ-12 score difference of 0.59 points (95% CI [0.32, 0.86]) and a 2.1 percentage point increase in the prevalence of psychological distress (p \u0026lt; 0.05). In contrast, there is no statistical evidence for survey effects on GHQ-12 scores among respondents aged 45 and above.\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eAverage Differences in Reported Mental Health Problems Between the UKHLS Main Survey and the COVID-19 Study by Age Group (N = 175,406).\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\"\u003e\u003cp\u003eAge Group\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\"\u003e\u003cp\u003eGHQ-12 (Score)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\"\u003e\u003cp\u003eGHQ-12 (Cases)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e16–29\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.796\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.173)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.021\u003c/p\u003e\u003cp\u003e(0.012)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e30–44\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.588\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.137)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.021\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.009)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e45–59\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.072\u003c/p\u003e\u003cp\u003e(0.103)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.009\u003c/p\u003e\u003cp\u003e(0.007)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"3\"\u003eNote: For each outcome, a survey-weighted regression was estimated with interactions among gender, age group, and COVID-19 study status, adjusting for the year and month of sample issuance. The average marginal effects of the COVID-19 study were computed by age group.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003cp\u003eStandard errors in parentheses. \u003csup\u003e*\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05, \u003csup\u003e**\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01, \u003csup\u003e***\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001\u003c/p\u003e\u003cp\u003eSource: UKHLS main waves 12 and 13, UKHLS COVID-19 study. Author’s calculations.\u003c/p\u003e\u003cp\u003eThese findings suggest that reported levels of psychological distress differ between the UKHLS main survey and the UKHLS COVID-19 study, with elevated distress among younger respondents and lower distress among older respondents in the COVID-19 study relative to the main survey. The observed differences could be driven by selection effects, survey design variations, or administrative differences. While these results do not imply that one survey provides a more accurate estimate of mental health, they underscore the importance of accounting for potential measurement differences. Failing to do so could lead to biased estimates when assessing changes in mental health before, during, and after the pandemic, particularly for younger age groups.\u003c/p\u003e\u003cp\u003eLong-Term Trends in Mental Health Since 2001\u003c/p\u003e\u003cp\u003eThis section examines long-term trends in mental health among 16-29-year-olds from 2001 onward, based on linear fixed effects estimations of Eq.\u0026nbsp;(\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). At this stage, the focus is on the underlying time trend \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:g\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e, which is modelled using year dummies. To isolate period trends from confounding life course effects, the model includes individual fixed effects, which adjust for cohort differences and three broad age groups (\u0026lt; 20, 20–24, and 25+), effectively constraining age effects. This effectively transforms Eq.\u0026nbsp;(\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e) into an age-period-cohort (APC) model (Fosse \u0026amp; Winship, \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e). Figure \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e displays the estimated change in psychological distress over time, adjusted for individual heterogeneity and life stage.\u003c/p\u003e\u003cp\u003eThe estimates indicate a steady and substantial increase in psychological distress among young adults. Between 2001 and 2019, the GHQ-12 score increased by 2.9 score points within individuals on average (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e, left plot), a difference that is both statistically and quantitatively significant. This increase represents 49% of the pooled standard deviation of the GHQ-12 scale (95% CI [41.5%, 57.5%]).\u003c/p\u003e\u003cp\u003eA similar trend is observed in the proportion of clinically relevant cases of psychological distress (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e, right panel). Compared to 2001, its prevalence had risen by 8.9 percentage points within individuals by 2019. In relative terms, young adults in 2019 were 39% more likely to meet the distress threshold compared to those in 2001 (95% CI: [18.5%, 59.3%]), an increase equivalent to approximately 1.1\u0026nbsp;million additional cases per year in the age group.\u003c/p\u003e\u003cp\u003eBoth plots peaked in 2020, followed by improvements, especially in cases of psychological distress. The following section will delve deeper into these patterns to unpick what changes were due to COVID-19 and the result of pre-existing trends.\u003c/p\u003e\u003cp\u003eFor context, the trend towards mental ill-health, with a subsequent flatting in the wake of the COVID-19 pandemic, correlated with averages in prescribed medication in mental health per 15–29-year-old per year in England since 2016 (GHQ-12 Score: r = 0.96, p \u0026lt; 0.001; GHQ-12 Cases: r = 0.79, p = 0.020). The long-term increase in psychological distress among young adults highlights the need to assess potential pandemic effects conditional on changes that might have happened irrespective of the pandemic. Ignoring the long-term trends towards mental ill-health can upward bias or overstate the estimated impact of the COVID-19 shock on mental health and understate the subsequent recovery.\u003c/p\u003e\u003cp\u003eCOVID-19 Impact: Deterioration and Recovery\u003c/p\u003e\u003cp\u003eTable \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e presents the estimated impact of the COVID-19 pandemic on psychological distress, based on Eq. (\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). Column (1) provides estimates from a baseline model that adjusts for time trends but does not account for survey differences between the UKHLS main survey and the COVID-19 study. Column (2) presents the headline findings, incorporating covariates to adjust for survey design effects. Column (3) restricts the analysis to UKHLS main survey data, allowing for an assessment of how well the dummy variable adjustment strategy accounts for survey differences. Column (4) examines the sensitivity of the headline results by excluding life-course markers from the model. The top panel of Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e reports results for the GHQ-12 score, while the bottom panel presents findings for GHQ-12 cases with clinically relevant levels of psychological distress.\u003c/p\u003e\u003cp\u003eColumn (1) estimates a 0.95-point increase in the GHQ-12 mean score and a 6 percentage-point increase in clinically relevant cases of psychological distress during the lockdown period relative to the expected trend. Levels of psychological distress returned to trend in the post-lockdown period. However, these estimates do not account for potential survey mode effects. Column (2) presents the headline findings, which incorporate adjustments for survey differences. The estimates represent the difference between observed mental ill-health and the counterfactual level of mental ill-health had pre-pandemic trends continued and in the absence of survey response effects. The GHQ-12 mean score was 0.54 points higher than expected during the lockdown, while post-lockdown levels were 0.15 points below expected trends, although this difference was not statistically significant.\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eMental Health During and After the COVID-19 Pandemic.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\"\u003e\u003cp\u003e(1)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\"\u003e\u003cp\u003e(2)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\"\u003e\u003cp\u003e(3)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\"\u003e\u003cp\u003e(4)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eTrend adjusted\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e+ Design adjusted\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eMainstage Only\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eW/o Covariates\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"4\"\u003e\u003cp\u003eGHQ-12 Score\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eLockdown\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.951\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.075)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.538\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.080)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.541\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.080)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.540\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.080)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003ePost-Lockdown\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.030\u003c/p\u003e\u003cp\u003e(0.087)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.150\u003c/p\u003e\u003cp\u003e(0.090)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.139\u003c/p\u003e\u003cp\u003e(0.091)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.150\u003c/p\u003e\u003cp\u003e(0.090)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eDifference\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.921\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.089)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.688\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.099)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.680\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.100)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.690\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.099)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"4\"\u003e\u003cp\u003eGHQ-12 Cases\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eLockdown\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.060\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.006)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.045\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.006)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.046\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.006)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e0.045\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.006)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003ePost-Lockdown\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.006\u003c/p\u003e\u003cp\u003e(0.007)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.013\u003c/p\u003e\u003cp\u003e(0.007)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.012\u003c/p\u003e\u003cp\u003e(0.007)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.013\u003c/p\u003e\u003cp\u003e(0.007)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eDifference\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.067\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.007)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.058\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.008)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.058\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.008)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e-0.058\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e(0.008)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eSeasonally adjusted\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eLife course controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eDesign\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eTime trend\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eObservations\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e96,564\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e96,564\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e84,745\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e96,564\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\"\u003e\u003cp\u003eIndividuals\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e18,027\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e18,027\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e17,985\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\"\u003e\u003cp\u003e18,027\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003eNote: Results from linear fixed effects regression models estimating Eq.\u0026nbsp;(\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e) for GHQ-12 scores and GHQ-12 cases. All models control for individual fixed effects, interview month, and a cubic polynomial in survey years since 2009. Column (2) adjusts for survey design effects using dummy variables for the COVID-19 study, interacted with interview year-quarter dummies. Column (3) restricts the sample to UKHLS main survey data. Column (4) excludes life-course markers (employment, partnership status, number of children under five).\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003cp\u003eStandard errors in parentheses. \u003csup\u003e*\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05, \u003csup\u003e**\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01, \u003csup\u003e***\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001\u003c/p\u003e\u003cp\u003eSource: UKHLS, UKHLS COVID-19 survey. Authors’ calculations.\u003c/p\u003e\u003cp\u003eThe increase in GHQ-12 scores during the lockdown period, while statistically significant, was quantitatively modest, amounting to 9.2% of the score’s standard deviation (95% CI: [6.5%, 11.8%]). Similarly, the prevalence of clinically relevant psychological distress was 4.5 percentage points above expected levels in the lockdown phase, implying that young adults were 19% more likely to meet the distress threshold (95% CI: [13.9%, 24.6%]). After March 2021, psychological distress recovered to expected levels, consistent with a return to the long-term trend. Taken together, these findings indicate that the direct impact of the COVID-19 pandemic on psychological distress was moderate and time-limited for the average 16-29-year-old with no apparent scarring of mental health from repeated lockdowns. The initial shock was followed by a recovery to trend levels.\u003c/p\u003e\u003cp\u003eColumn (3) removes observations from the COVID-19 study, leaving only UKHLS main survey data. The similarity between columns (2) and (3) suggests that the dummy variable adjustment approach was sufficient to adjust for survey response effects on mental health reporting. Column (4) removes life-course markers from the model, with minimal impact on the headline findings. This suggests that the pandemic effect did not operate through partnership and family formation.\u003c/p\u003e\u003cp\u003eOnce long-term trends and survey response effects are accounted for, the findings suggest that the mental health impact of the COVID-19 shock among young adults was limited in duration, with levels returning to expected trends after March 2021. Despite the substantial disruptions experienced during the pandemic, these findings indicate that, on average, young adults did not experience lasting scarring effects. However, this temporary setback and subsequent recovery must be understood within the broader context of a sustained long-term rise in psychological distress, which preceded the pandemic.\u003c/p\u003e\u003cp\u003eCalendar Quarter Changes in Psychological Distress\u003c/p\u003e\u003cp\u003eThe broadly defined lockdown and post-lockdown periods may have smoothed out short-term shifts in psychological distress during the pandemic. To capture these variations, this section re-estimates Eq.\u0026nbsp;(\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e) using the preferred specification from Column (2) of Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e, detailing changes in mental health by calendar quarter from the first quarter of 2020 (2020Q1) to the fourth quarter of 2021 (2021Q4), when the COVID-19 survey provided an expanded sample. Figure \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e presents these results, illustrating quarterly changes in GHQ-12 scores and the prevalence of psychological distress relative to their counterfactual trends.\u003c/p\u003e\u003cp\u003eThe GHQ-12 score was .43 points (95% CI: 0.17; 0.69), and cases of psychological distress were 5.4 percentage points (95% CI: 3.2; 7.6) above-trend in 2020Q2, during the first UK lockdown. In relative terms, clinically relevant cases of psychological distress exceeded their counterfactual prediction by 14.7%, while the GHQ-12 rose + 7.3% standard deviations. A slight decline in psychological distress followed before another rise over the winter of 2020/2021, coinciding with renewed lockdown measures. The pandemic effect on GHQ-12 scores peaked in 2021Q1 with 0.9 score points above the counterfactual estimate, while the prevalence of GHQ-12 cases was comparable to its level during the initial pandemic shock.\u003c/p\u003e\u003cp\u003eMental health indicators improved after 2021Q1. By the second quarter of 2021 (2021Q2), reported distress levels were statistically indistinguishable from the counterfactual trend. This suggests that the direct effect of the COVID-19 pandemic on psychological distress was concentrated in the periods of stringent lockdown measures. The full return of psychological distress levels to the predicted counterfactual trend also supports the parallel trends assumption. This finding suggests that, in the absence of the COVID-19 shock, mental health outcomes would have continued along their pre-pandemic trajectory, reinforcing the credibility of the identification strategy.\u003c/p\u003e\u003cp\u003eThe results confirm that the impact of the pandemic on young adults' mental health was most pronounced in periods of widespread restrictions, with no evidence of persistent mental health deterioration at the population level beyond that point. These results are consistent with the broader trend analysis, which indicates that the rise in psychological distress associated with COVID-19 was temporary and that distress levels returned to their pre-pandemic trajectory once restrictions were permanently lifted, starting in 2021Q2.\u003c/p\u003e\u003cp\u003eHeterogeneous Effects\u003c/p\u003e\u003cp\u003eThe mental health impact of the COVID-19 pandemic may have varied across demographic and socioeconomic groups due to, for example, differences in resources, behaviour, typical activities, or the level of disruptions.\u003c/p\u003e\u003cp\u003eHowever, re-estimating Eq.\u0026nbsp;(\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e), including adjustments for survey response effects and time trends in pooled panel samples of 16–59-year-olds, does not suggest notable differences in the mental health response by age over the lockdown and post-lockdown period, as depicted in Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e. On average across age groups, GHQ-12 was 0.46 points (95% CI [0.36; 0.56]) higher than expected during the acute phase and − 0.11 score points (95% CI [− .21; − .004]) below trend in the recovery phase. Similarly, clinically relevant cases of psychological distress were 4.0 percentage points (95% CI: 3.2; 4.8) above their counterfactual prediction in the acute phase and indistinguishable from trend thereafter (-0.005, 95% CI: − .013; 0.003). Wald tests fail to reject the null hypotheses of age-homogenous mental health responses in response to the lockdown and post-lockdown periods across age groups (GHQ-12 score, p = 0.757; GHQ-12 cases, p = 0.147).\u003c/p\u003e\u003cp\u003eWithin the sample of young adults, we also tested for differences in COVID-19 effects across gender (male/ female), age (16–21, 22–29), ethnicity (white/ ethnic minorities), baseline household income rank (bottom two-thirds, top third), and economic activity (not in work/ in work). The results indicate some variation in the extent to which different subgroups experienced distress during the lockdown period of the pandemic (see Tables \u003cspan class=\"InternalRef\"\u003eA1\u003c/span\u003e and \u003cspan class=\"InternalRef\"\u003eA2\u003c/span\u003e in the supplement).\u003c/p\u003e\u003cp\u003eFor GHQ-12 scores, we note a significantly higher psychological distress effect of the lockdown period on young women and respondents in the top third of the household income distribution than the rest. In contrast, young people in work and ethnic minorities experienced a lower effect of the lockdown period on psychological distress than their counterparts (see Table \u003cspan class=\"InternalRef\"\u003eA1\u003c/span\u003e in the supplement); although in the case of ethnic minority groups, the difference did not reach the 5% level of significance. All subgroups’ GHQ-12 sum scores were statistically indistinguishable from their counterfactual trend post-lockdown. Findings for GHQ-12 cases confirm patterns of heterogeneity by gender (women were hit harder) and ethnic minority, but not by economic activity (Table \u003cspan class=\"InternalRef\"\u003eA2\u003c/span\u003e in the supplement).\u003c/p\u003e\u003cp\u003eAdditional Analyses and Robustness Checks\u003c/p\u003e\u003cp\u003eWe assessed COVID-19’s effect on life satisfaction and reported loneliness using Eq.\u0026nbsp;(\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e), finding minimally higher-than-expected life satisfaction and lower-than-expected loneliness during the post-lockdown period. There is no evidence that either was adversely affected by the COVID-19 pandemic during lockdown (Table \u003cspan class=\"InternalRef\"\u003eA3\u003c/span\u003e in the supplement reports). The estimates confirm the improvements in psychological well-being with receding feelings of loneliness post-lockdown.\u003c/p\u003e\u003cp\u003eIt is conceivable that the relatively small mental health effects are due to individuals in distress selecting out of the study. Like any panel study, UKHLS suffers from attrition over time. If this is the case, we might underestimate the impact of COVID-19 on mental ill-health. Therefore, we added indicator variables measuring next-wave retention – a respondent participated in the following wave and previous-wave retention – a respondent had participated in the previous wave – in turn to Eq.\u0026nbsp;(\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). These tests were conducted using both continuous psychological distress scores (GHQ-12 score) and clinically relevant cases of distress (GHQ-12 cases). The results for all tests were statistically insignificant (Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e), meaning that panel retention did not predict psychological distress. These results do not suggest that panel retention was systematically related to mental health status.\u003c/p\u003e\u003cp\u003eTable 4: \u003cem\u003eAdded Variable Test for Panel Retention.\u003c/em\u003e\u003c/p\u003e\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd valign=\"top\" style=\"width: 200px;\"\u003e\u003cp\u003e\u0026nbsp;\u003c/p\u003e\u003c/td\u003e\u003ctd valign=\"top\" style=\"width: 200px;\"\u003e\u003cp\u003eGHQ-12 scores\u003c/p\u003e\u003c/td\u003e\u003ctd valign=\"top\" style=\"width: 200px;\"\u003e\u003cp\u003eGHQ-12 cases\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd valign=\"top\" style=\"width: 200px;\"\u003e\u003cp\u003eNext-wave retention\u0026nbsp;\u003c/p\u003e\u003c/td\u003e\u003ctd valign=\"top\" style=\"width: 200px;\"\u003e\u003cp\u003eF(1, 14015) = 0.01, p = 0.930\u003c/p\u003e\u003c/td\u003e\u003ctd valign=\"top\" style=\"width: 200px;\"\u003e\u003cp\u003eF(1,14015) = 0.01, p = 0.907\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd valign=\"top\" style=\"width: 200px;\"\u003e\u003cp\u003ePrevious-wave retention\u0026nbsp;\u003c/p\u003e\u003c/td\u003e\u003ctd valign=\"top\" style=\"width: 200px;\"\u003e\u003cp\u003eF(1, 14015) = 0.13, p = 0.721\u003c/p\u003e\u003c/td\u003e\u003ctd valign=\"top\" style=\"width: 200px;\"\u003e\u003cp\u003eF(1,14015) = 0.17, p = 0.679\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003cdiv class=\"gridtable\"\u003e\u003cp\u003eNote: F test results from variables measuring next-wave / previous-wave retention added to Equation (1)\u003c/p\u003e\u003cp\u003eSource: UKHLS, UKHLS COVID-19 Study. Author’s calculations.\u003c/p\u003e\u003cp\u003eOur identification of COVID-19-related mental health consequences relies on an appropriately fitted time trend, which can be difficult, especially towards the endpoints where data is sparse. To assess the robustness of the findings against an alternative specification, we re-estimated Equation (1) with a restricted cubic spline in the interview date with five knots. Spline functions are piecewise-defined polynomials that are combined in such a way that they are smooth at the points where the pieces join, called knots. They are used in regression analysis to model nonlinear relationships between variables (Perperoglou et al., 2019). A restricted cubic spline is linear before the first knot and after the last knot, which can help extrapolation. The estimates suggest 0.50 points (p \u0026lt; 0.001) higher than expected GHQ-12 score on average during the lockdown phase and a score of -0.14 (p = 0.125) statistically indistinct from trend after that. The prevalence of clinically relevant psychological distress was raised by 4.3 percentage points (p \u0026lt; 0.001) compared to the counterfactual prediction in the lockdown period and statistically indistinct from its trend value afterwards (-.013, p = 0.07), confirming the modest and time-limited effect of the pandemic on young people’s mental health as measured by the GHQ-12 instrument.\u003c/p\u003e\u003cp\u003eFinally, to ascertain how well the approach is able to separate out shocks from trends, we conducted a placebo test whereby we ‘switched on’ dummies for the more ‘normal’ years 2017 (Grenfell Tower Fire, Corbyn’s defeat in the General Election, and #MeToo UK) and 2018 (Windrush scandal, royal wedding between Harry and Meghan, and arrival of TikTok), instead of the lockdown and post-lockdown dummy. The results for pseudo-treatments were statistically insignificant at common levels (Table A4), supporting the validity of the parallel trends assumption. \u0026nbsp;\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion and Conclusion","content":"\u003ch2\u003eDiscussion\u003c/h2\u003e\n\u003cp\u003eThis study examines COVID-19\u0026rsquo;s short and long-term mental health impact on 16-29-year-olds in the United Kingdom against longer-term trends, drawing on nationally representative longitudinal data from 2001 to 2023. The data enables a focus on pre-existing trends and experiences during and after the COVID-19 pandemic, controlling for variations in mental health over the life course and potential survey response effects between different survey sources during the pandemic. The findings suggest that already before the COVID-19 pandemic, there had been a rise in mental distress among young people. COVID-19 temporarily accelerated this trend towards mental ill-health, followed by a recovery, with no adverse long-term pandemic-related impacts observed on average, consistent with receding feelings of loneliness and above-trend life satisfaction post-lockdowns. The results apply after adjusting for time-constant individual differences, composition, a non-linear time trend, seasonality, and survey response effects between the UKHLS main and COVID-19 surveys employing linear fixed effects regression models.\u003c/p\u003e \u003cp\u003eThe study provides much-needed evidence on the longer-term mental health trends among young adults in times of global upheavals. The increased levels of mental distress during the acute phase of the pandemic accelerated pre-existing trends, marked by the 2008 Great Recession, the subsequent UK government austerity programme, and an initial shock reaction in 2020 to the COVID-19 pandemic and lockdown measures, which then subsided after March 2021. Future research has yet to unpack the underlying changes driving the deterioration in young adults' mental health, which may include increases in social media use (Blanchflower, Bryson, et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), declining career outlooks, reduced real income, as well as new living arrangements with parents, partners, and others (Gagn\u0026eacute; et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), but also cuts to government spending for transport and youth services (Brown et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFocusing on experiences from 2020 to 2022, the findings support other studies that show that UK young adults experienced historically high levels of distress during the first national lockdown between April and July 2020. However, the findings suggest the initial mental health shock was i) smaller than often reported and ii) temporary after adjusting for pre-existing trends and different response patterns between data sources. The pandemic impact on mental health was stronger among women and young adults in the top third of the household income distribution and less impactful for ethnic minorities, echoing findings by Miall et al. (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) for children in the UK. There was no evidence for a stronger pandemic-related mental health decline among young than older adults above 30 years. Similarly, we find no adverse long-term consequences of the pandemic on life satisfaction and reported loneliness on average. The moderate, temporary COVID-19-related setback in mental health is consistent with findings for youths in Norway (Koz\u0026aacute;k et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) or European adults, more generally (Blanchflower \u0026amp; Bryson, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Heterogeneity in the mental health response within young adults during the COVID-19 pandemic might be due to differences in usual social interactions, spare time activities, participation in education and/or employment, or familiarity with stress and uncertainty. However, post-lockdown, average levels of psychological distress were back to their counterfactual trend.\u003c/p\u003e \u003cp\u003eThe results highlight that, on average, young adults have had the capacity to adapt to the upheaval of the COVID-19 pandemic. However, some adverse COVID-related experiences may continue to predict mental health problems beyond the short and medium term, underscoring the importance of considering individual-specific experiences (Anders \u0026amp; Holt-White, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Furthermore, the analysis highlights the magnitude of the mental health burden already present in the UK\u0026rsquo;s young adult population in the years leading up to 2020. Evidence of resilience during the pandemic suggests that while immediate support and interventions during the lockdown phases of crises were critical, long-term policies should focus on addressing the underlying deterioration in mental health among young adults. Strengthening mental health services and support systems and addressing the root causes to tackle the long-term rise in mental health problems is crucial for improving overall well-being and functioning.\u003c/p\u003e \u003cp\u003eStrengths and limitations\u003c/p\u003e \u003cp\u003eThe study combines nationally representative longitudinal data, including a validated measure of mental health, with a plausible strategy to estimate the causal effects of the COVID-19 pandemic in the short and long term. However, despite its strengths, there are a few possible limitations. First, while the observed trend compares well with patterns in prescriptions for medicines used in mental health, self-reported measures of psychological distress may introduce reporting biases. Second, the estimation of long-term mental health trends relies on a stepwise, coarsened specification of age to remain identifiable. This simplification could introduce bias. Third, while the study adjusts for pre-pandemic trends, life stages, and individual fixed and survey response effects, other unobserved factors could influence mental health outcomes. Fourth, while there is no immediate evidence for violation of the parallel trends assumption that informs the interpretation of the mode parameters as ATT, any unaccounted deviation from the counterfactual trend that happened at the same time as the country entered the lockdown and post-lockdown phase might introduce bias. However, factors such as anticipation effects, seem unlikely and are not borne out in the data (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Fifth, the study does not disentangle the direct effect of the pandemic (e.g., mortality and morbidity risks) from indirect effects stemming from containment policies. Finally, attrition across UKHLS has been high among young adults, and the COVID-19 survey waves had relatively low response rates and small young adult samples, introducing selection bias and the threat of underpowered subgroup comparisons. The sensitivity tests, including checks for panel attrition, support the robustness of the findings but cannot entirely eliminate these limitations.\u003c/p\u003e\n\u003ch3\u003eConclusion\u003c/h3\u003e\n\u003cp\u003eDespite these limitations, this study offers unique evidence of trends in mental distress among young people in the UK over the past twenty-one years, including before, during and after the COVID-19 pandemic. Rising mental health problems among youth were already observed before the pandemic. We highlighted the increase in distress that young people have faced over time. Going back to trend levels of mental health should, therefore, not suffice as a public health target. The findings emphasise the need for systemic efforts to address the mental health problems among young adults and efforts to promote their well-being in the long term. Relevant initiatives must consider that despite the recovery following the COVID-19 pandemic, pockets of increased vulnerabilities might compound into potential scarring effects on future outcomes.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData Sharing and Ethical Approval\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data used in this study are available from the UK Data Service (study numbers 6614 and 8644). The University of Essex Ethics Committee granted ethical approval for data collection. Syntax replicating analyses, tables and figures is available at https://doi.org/10.5522/04/28469006.v1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding \u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThere was no funding source for this study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe declare no competing interests.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis paper was made possible by data collection funded by the UK Research and Innovation via their COVID response fund and the Economic and Social Research Council. We are grateful for Lulei (Shirley) Chen\u0026rsquo;s contribution to an earlier draft.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAhmed, N., Barnett, P., Greenburgh, A., Pemovska, T., Stefanidou, T., Lyons, N., Ikhtabi, S., Talwar, S., Francis, E. R., Harris, S. M., Shah, P., Machin, K., Jeffreys, S., Mitchell, L., Lynch, C., Foye, U., Schlief, M., Appleton, R., Saunders, K. R. K., . . . Johnson, S. (2023). Mental health in Europe during the COVID-19 pandemic: a systematic review. \u003cem\u003eLancet Psychiatry\u003c/em\u003e,\u003cem\u003e 10\u003c/em\u003e(7), 537-556. https://doi.org/10.1016/S2215-0366(23)00113-X\u003c/li\u003e\n\u003cli\u003eAnaya, L., Howley, P., Waqas, M., \u0026amp; Yalonetzky, G. (2023). Locked down in distress: A quasi‐experimental estimation of the mental‐health fallout from the COVID‐19 pandemic. \u003cem\u003eEconomic Inquiry\u003c/em\u003e,\u003cem\u003e 62\u003c/em\u003e(1), 56-73. https://doi.org/10.1111/ecin.13181\u003c/li\u003e\n\u003cli\u003eAnders, J., \u0026amp; Holt-White, E. (2024). Young people\u0026rsquo;s subjective wellbeing in the wake of the COVID-19 pandemic: evidence from a representative cohort study in England. \u003cem\u003eCEPEO Working Paper Series\u003c/em\u003e,\u003cem\u003e 24-05\u003c/em\u003e. https://ideas.repec.org/p/ucl/cepeow/24-05.html (UCL Centre for Education Policy \u0026amp; Equalising Opportunities)\u003c/li\u003e\n\u003cli\u003eAnyaegbu, G. (2010). Using the OECD equivalence scale in taxes and benefits analysis. \u003cem\u003eEconomic \u0026amp; Labour Market Review\u003c/em\u003e,\u003cem\u003e 4\u003c/em\u003e(1), 49-54. https://doi.org/10.1057/elmr.2010.9\u003c/li\u003e\n\u003cli\u003eBaksheev, G. N., Robinson, J., Cosgrave, E. M., Baker, K., \u0026amp; Yung, A. R. (2011). Validity of the 12-item General Health Questionnaire (GHQ-12) in detecting depressive and anxiety disorders among high school students. \u003cem\u003ePsychiatry Res\u003c/em\u003e,\u003cem\u003e 187\u003c/em\u003e(1-2), 291-296. https://doi.org/10.1016/j.psychres.2010.10.010\u003c/li\u003e\n\u003cli\u003eBanks, J., \u0026amp; Xu, X. (2020). The Mental Health Effects of the First Two Months of Lockdown during the COVID‐19 Pandemic in the UK. \u003cem\u003eFiscal Studies\u003c/em\u003e,\u003cem\u003e 41\u003c/em\u003e(3), 685-708. https://doi.org/10.1111/1475-5890.12239\u003c/li\u003e\n\u003cli\u003eBlanchflower, D. G. (2025). Declining Youth Well-being in 167 UN Countries. Does Survey Mode, or Question Matter? \u003cem\u003eNBER WORKING PAPER\u003c/em\u003e(33415). https://doi.org/10.3386/w33415\u003c/li\u003e\n\u003cli\u003eBlanchflower, D. G., \u0026amp; Bryson, A. (2024). Were COVID and the Great Recession well-being reducing? \u003cem\u003ePLOS ONE\u003c/em\u003e,\u003cem\u003e 19\u003c/em\u003e(11), e0305347. https://doi.org/10.1371/journal.pone.0305347\u003c/li\u003e\n\u003cli\u003eBlanchflower, D. G., Bryson, A., Lepinteur, A., \u0026amp; Piper, A. (2024). Further Evidence on the Global Decline in the Mental Health of the Young. \u003cem\u003eNBER WORKING PAPER\u003c/em\u003e(32500). https://doi.org/10.3386/w32500\u003c/li\u003e\n\u003cli\u003eBlanchflower, D. G., Bryson, A., \u0026amp; Xu, X. (2024). The Declining Mental Health Of The Young And The Global Disappearance Of The Hump Shape In Age In Unhappiness. \u003cem\u003eNBER WORKING PAPER\u003c/em\u003e. https://doi.org/10.3386/w32337\u003c/li\u003e\n\u003cli\u003eBrown, H., Gao, N., \u0026amp; Song, W. (2024). Regional trends in mental health inequalities in young people aged 16\u0026ndash;25 in the UK and the role of cuts to local government expenditure: Repeated cross-sectional analysis using the British household panel Survey/UK household longitudinal survey. \u003cem\u003eSocial Science \u0026amp; Medicine\u003c/em\u003e,\u003cem\u003e 353\u003c/em\u003e, 117068. https://doi.org/10.1016/j.socscimed.2024.117068\u003c/li\u003e\n\u003cli\u003eBurton, J., Lynn, P., \u0026amp; Benzeval, M. (2020). How understanding society: the UK household longitudinal study adapted to the COVID-19 pandemic. \u003cem\u003eSurvey Research Methods\u003c/em\u003e,\u003cem\u003e 14\u003c/em\u003e(2), 235-239. https://doi.org/10.18148/srm/2020.v14i2.7746\u003c/li\u003e\n\u003cli\u003eChaudhuri, K., \u0026amp; Howley, P. (2022). The impact of COVID-19 vaccination for mental well-being. \u003cem\u003eEuropean Economic Review\u003c/em\u003e,\u003cem\u003e 150\u003c/em\u003e, 104293. https://doi.org/10.1016/j.euroecorev.2022.104293\u003c/li\u003e\n\u003cli\u003eConti, G., \u0026amp; Pudney, S. (2011). Survey Design and the Analysis of Satisfaction. \u003cem\u003eThe Review of Economics and Statistics\u003c/em\u003e,\u003cem\u003e 93\u003c/em\u003e(3). https://doi.org/10.1162/REST_a_00202\u003c/li\u003e\n\u003cli\u003eDavillas, A., de Oliveira, V. H., \u0026amp; Jones, A. M. (2023). Is inconsistent reporting of self-assessed health persistent and systematic? Evidence from the UKHLS. \u003cem\u003eEconomics \u0026amp; Human Biology\u003c/em\u003e,\u003cem\u003e 49\u003c/em\u003e, 101219. https://doi.org/10.1016/j.ehb.2022.101219\u003c/li\u003e\n\u003cli\u003eDhensa-Kahlon, R. K., Wan, S. T., Coyle-Shapiro, J. A.-M., \u0026amp; Teoh, K. R.-H. (2025). The mental health impact of repeated COVID-19 enforced lockdowns in England: evidence from the UK Household Longitudinal Study. \u003cem\u003eBJPsych Open\u003c/em\u003e,\u003cem\u003e 11\u003c/em\u003e(1). https://doi.org/10.1192/bjo.2024.803\u003c/li\u003e\n\u003cli\u003eDuarte Neves, H., Asaria, M., \u0026amp; Stabile, M. (2024). Young, Muslim and poor: The persistent impacts of the pandemic on mental health in the UK. \u003cem\u003eSocial Science \u0026amp; Medicine\u003c/em\u003e,\u003cem\u003e 353\u003c/em\u003e, 117032. https://doi.org/10.1016/j.socscimed.2024.117032\u003c/li\u003e\n\u003cli\u003eFancourt, D., Steptoe, A., \u0026amp; Bu, F. (2021). Trajectories of anxiety and depressive symptoms during enforced isolation due to COVID-19 in England: a longitudinal observational study. \u003cem\u003eThe Lancet Psychiatry\u003c/em\u003e,\u003cem\u003e 8\u003c/em\u003e(2). https://doi.org/10.1016/S2215-0366(20)30482-X\u003c/li\u003e\n\u003cli\u003eFosse, E., \u0026amp; Winship, C. (2019). Analyzing Age-Period-Cohort Data: A Review and Critique. \u003cem\u003eAnnual Review of Sociology\u003c/em\u003e,\u003cem\u003e 45\u003c/em\u003e(Volume 45, 2019), 467-492. https://doi.org/https://doi.org/10.1146/annurev-soc-073018-022616\u003c/li\u003e\n\u003cli\u003eGagn\u0026eacute;, T., Sacker, A., \u0026amp; Schoon, I. (2022). Transition milestones and life satisfaction at ages 25/26 among cohorts born in 1970 and 1989\u0026ndash;90. \u003cem\u003eAdvances in Life Course Research\u003c/em\u003e,\u003cem\u003e 51\u003c/em\u003e, 100463. https://doi.org/10.1016/j.alcr.2022.100463\u003c/li\u003e\n\u003cli\u003eGagn\u0026eacute;, T., Schoon, I., McMunn, A., Sacker, A., Gagn\u0026eacute;, T., Schoon, I., McMunn, A., \u0026amp; Sacker, A. (2021). Mental distress among young adults in Great Britain: long-term trends and early changes during the COVID-19 pandemic. \u003cem\u003eSocial Psychiatry and Psychiatric Epidemiology 2021 57:6\u003c/em\u003e,\u003cem\u003e 57\u003c/em\u003e(6). https://doi.org/10.1007/s00127-021-02194-7\u003c/li\u003e\n\u003cli\u003eHenseke, G., Green, F., Schoon, I., Henseke, G., Green, F., \u0026amp; Schoon, I. (2022). Living with COVID-19: Subjective Well-Being in the Second Phase of the Pandemic. \u003cem\u003eJournal of Youth and Adolescence 2022 51:9\u003c/em\u003e,\u003cem\u003e 51\u003c/em\u003e(9). https://doi.org/10.1007/s10964-022-01648-8\u003c/li\u003e\n\u003cli\u003eHern\u0026aacute;n, M. A. (2018). The C-Word: Scientific Euphemisms Do Not Improve Causal Inference From Observational Data. \u003cem\u003eAmerican Journal of Public Health\u003c/em\u003e,\u003cem\u003e 108\u003c/em\u003e(5), 616-619. https://doi.org/10.2105/ajph.2018.304337\u003c/li\u003e\n\u003cli\u003eHotopf, M., Bullmore, E., O\u0026apos;Connor, R. C., \u0026amp; Holmes, E. A. (2020). The scope of mental health research during the COVID-19 pandemic and its aftermath. \u003cem\u003eThe British Journal of Psychiatry\u003c/em\u003e,\u003cem\u003e 217\u003c/em\u003e(4), 540-542. https://doi.org/10.1192/bjp.2020.125\u003c/li\u003e\n\u003cli\u003eInstitute for Social and Economic Research. (2021a). Understanding Society COVID-19 User Guide. Version 10.0. In. Colchester: University of Essex.\u003c/li\u003e\n\u003cli\u003eInstitute for Social and Economic Research. (2021b). \u003cem\u003eUnderstanding Society: COVID-19 Study, 2020-2021\u003c/em\u003e (8644; Version 11th Edition). https://doi.org/10.5255/UKDA-SN-8644-11\u003c/li\u003e\n\u003cli\u003eInstitute for Social and Economic Research. (2023). \u003cem\u003eUnderstanding Society. [data series].\u003c/em\u003e (2000053; Version 11th Release) UK Data Service. https://doi.org/10.5255/UKDA-Series-2000053\u003c/li\u003e\n\u003cli\u003eJaschke, P., Kosyakova, Y., Kuche, C., Walther, L., Go\u0026szlig;ner, L., Jacobsen, J., Ta, T. M. T., Hahn, E., Hans, S., \u0026amp; Bajbouj, M. (2023). Mental health and well-being in the first year of the COVID-19 pandemic among different population subgroups: evidence from representative longitudinal data in Germany. \u003cem\u003eBMJ Open\u003c/em\u003e,\u003cem\u003e 13\u003c/em\u003e(6). https://doi.org/10.1136/bmjopen-2022-071331 \u003c/li\u003e\n\u003cli\u003eKoz\u0026aacute;k, M., Bakken, A., von Soest, T., Koz\u0026aacute;k, M., Bakken, A., \u0026amp; von Soest, T. (2023). Psychosocial well-being before, during and after the COVID-19 pandemic: a nationwide study of more than half a million Norwegian adolescents. \u003cem\u003eNature Mental Health 2023 1:7\u003c/em\u003e,\u003cem\u003e 1\u003c/em\u003e(7). https://doi.org/10.1038/s44220-023-00088-y\u003c/li\u003e\n\u003cli\u003eKung, C. S. J., Kunz, J. S., \u0026amp; Shields, M. A. (2023). COVID-19 lockdowns and changes in loneliness among young people in the U.K. \u003cem\u003eSocial Science \u0026amp; Medicine\u003c/em\u003e,\u003cem\u003e 320\u003c/em\u003e, 115692. https://doi.org/10.1016/j.socscimed.2023.115692\u003c/li\u003e\n\u003cli\u003eLundberg, I., Johnson, R., \u0026amp; Stewart, B. M. (2021). What Is Your Estimand? Defining the Target Quantity Connects Statistical Evidence to Theory. \u003cem\u003eAmerican Sociological Review\u003c/em\u003e,\u003cem\u003e 86\u003c/em\u003e(3), 532-565. https://doi.org/10.1177/00031224211004187\u003c/li\u003e\n\u003cli\u003eLundin, A., Hallgren, M., Theobald, H., Hellgren, C., \u0026amp; Torgen, M. (2016). Validity of the 12-item version of the General Health Questionnaire in detecting depression in the general population. \u003cem\u003ePublic Health\u003c/em\u003e,\u003cem\u003e 136\u003c/em\u003e, 66-74. https://doi.org/10.1016/j.puhe.2016.03.005\u003c/li\u003e\n\u003cli\u003eMcManus, S., Gunnell, D., McManus, S., \u0026amp; Gunnell, D. (2019). Trends in mental health, non‐suicidal self‐harm and suicide attempts in 16\u0026ndash;24-year old students and non-students in England, 2000\u0026ndash;2014. \u003cem\u003eSocial Psychiatry and Psychiatric Epidemiology 2019 55:1\u003c/em\u003e,\u003cem\u003e 55\u003c/em\u003e(1). https://doi.org/10.1007/s00127-019-01797-5\u003c/li\u003e\n\u003cli\u003eMiall, N., Pearce, A., Moore, J. C., Benzeval, M., \u0026amp; Green, M. J. (2023). Inequalities in children\u0026rsquo;s mental health before and during the COVID-19 pandemic: findings from the UK Household Longitudinal Study. \u003cem\u003eJ Epidemiol Community Health\u003c/em\u003e,\u003cem\u003e 77\u003c/em\u003e(12), 762-769. https://doi.org/10.1136/jech-2022-220188\u003c/li\u003e\n\u003cli\u003eNeugebauer, M., Patzina, A., Dietrich, H., \u0026amp; Sandner, M. (2023). Two pandemic years greatly reduced young people\u0026rsquo;s life satisfaction: evidence from a comparison with pre-COVID-19 panel data. \u003cem\u003eEuropean Sociological Review\u003c/em\u003e. https://doi.org/10.1093/esr/jcad077\u003c/li\u003e\n\u003cli\u003eNiedzwiedz, C. L., Green, M. J., Benzeval, M., Campbell, D., Craig, P., Demou, E., Leyland, A., Pearce, A., Thomson, R., Whitley, E., \u0026amp; Katikireddi, S. V. (2021). Mental health and health behaviours before and during the initial phase of the COVID-19 lockdown: longitudinal analyses of the UK Household Longitudinal Study. \u003cem\u003eJ Epidemiol Community Health\u003c/em\u003e,\u003cem\u003e 75\u003c/em\u003e(3). https://doi.org/10.1136/jech-2020-215060\u003c/li\u003e\n\u003cli\u003eO\u0026apos;Connor, R. C., Wetherall, K., Cleare, S., McClelland, H., Melson, A. J., Niedzwiedz, C. L., O\u0026apos;Carroll, R. E., O\u0026apos;Connor, D. B., Platt, S., Scowcroft, E., Watson, B., Zortea, T., Ferguson, E., \u0026amp; Robb, K. A. (2021). Mental health and well-being during the COVID-19 pandemic: longitudinal analyses of adults in the UK COVID-19 Mental Health \u0026amp; Wellbeing study | The British Journal of Psychiatry | Cambridge Core. \u003cem\u003eThe British Journal of Psychiatry\u003c/em\u003e,\u003cem\u003e 218\u003c/em\u003e(6). https://doi.org/10.1192/bjp.2020.212\u003c/li\u003e\n\u003cli\u003ePerperoglou, A., Sauerbrei, W., Abrahamowicz, M., \u0026amp; Schmid, M. (2019). A review of spline function procedures in R. \u003cem\u003eBMC Medical Research Methodology\u003c/em\u003e,\u003cem\u003e 19\u003c/em\u003e(1), 46. https://doi.org/10.1186/s12874-019-0666-3\u003c/li\u003e\n\u003cli\u003ePevalin, D. J. (2000). Multiple applications of the GHQ-12 in a general population sample: an investigation of long-term retest effects. \u003cem\u003eSocial Psychiatry and Psychiatric Epidemiology\u003c/em\u003e,\u003cem\u003e 35\u003c/em\u003e(11), 508-512. https://doi.org/10.1007/s001270050272\u003c/li\u003e\n\u003cli\u003ePierce, M., Hope, H., Ford, T., Hatch, S., Hotopf, M., John, A., Kontopantelis, E., Webb, R., Wessely, S., McManus, S., \u0026amp; Abel, K. M. (2020). Mental health before and during the COVID-19 pandemic: a longitudinal probability sample survey of the UK population. \u003cem\u003eLancet Psychiatry\u003c/em\u003e,\u003cem\u003e 7\u003c/em\u003e(10), 883-892. https://doi.org/10.1016/S2215-0366(20)30308-4\u003c/li\u003e\n\u003cli\u003eSerrano-Alarc\u0026oacute;n, M., Kentikelenis, A., Mckee, M., \u0026amp; Stuckler, D. (2022). Impact of COVID‐19 lockdowns on mental health: Evidence from a quasi‐natural experiment in England and Scotland. \u003cem\u003eHealth Economics\u003c/em\u003e,\u003cem\u003e 31\u003c/em\u003e(2). https://doi.org/10.1002/hec.4453\u003c/li\u003e\n\u003cli\u003eSettersten, R. A., Jr., Bernardi, L., Harkonen, J., Antonucci, T. C., Dykstra, P. A., Heckhausen, J., Kuh, D., Mayer, K. U., Moen, P., Mortimer, J. T., Mulder, C. H., Smeeding, T. M., van der Lippe, T., Hagestad, G. O., Kohli, M., Levy, R., Schoon, I., \u0026amp; Thomson, E. (2020). Understanding the effects of Covid-19 through a life course lens. \u003cem\u003eAdvances in Life Course Research\u003c/em\u003e,\u003cem\u003e 45\u003c/em\u003e, 100360. https://doi.org/10.1016/j.alcr.2020.100360\u003c/li\u003e\n\u003cli\u003eSun, Y., Wu, Y., Fan, S., Dal Santo, T., Li, L., Jiang, X., Li, K., Wang, Y., Tasleem, A., Krishnan, A., He, C., Bonardi, O., Boruff, J. T., Rice, D. B., Markham, S., Levis, B., Azar, M., Thombs-Vite, I., Neupane, D., . . . Thombs, B. D. (2023). Comparison of mental health symptoms before and during the covid-19 pandemic: evidence from a systematic review and meta-analysis of 134 cohorts. \u003cem\u003eBMJ\u003c/em\u003e,\u003cem\u003e 380\u003c/em\u003e, e074224. https://doi.org/10.1136/bmj-2022-074224\u003c/li\u003e\n\u003cli\u003eWright, N., Hill, J., Sharp, H., Refberg-Brown, M., Crook, D., Kehl, S., Pickles, A., Wright, N., Hill, J., Sharp, H., Refberg-Brown, M., Crook, D., Kehl, S., \u0026amp; Pickles, A. (2024). COVID-19 pandemic impact on adolescent mental health: a reassessment accounting for development. \u003cem\u003eEuropean Child \u0026amp; Adolescent Psychiatry\u003c/em\u003e. https://doi.org/10.1007/s00787-023-02337-y\u003c/li\u003e\n\u003cli\u003eZhang, A., Gagne, T., Walsh, D., Ciancio, A., Proto, E., \u0026amp; McCartney, G. (2023). Trends in psychological distress in Great Britain, 1991-2019: evidence from three representative surveys. \u003cem\u003eJ Epidemiol Community Health\u003c/em\u003e,\u003cem\u003e 77\u003c/em\u003e(7), 468-473. https://doi.org/10.1136/jech-2022-219660 \u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Footnotes","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e There were 167 interviews with the target age-group in 2024, which were assigned to their sampling year-month.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Psychological distress, COVID-19, Young Adults, Longitudinal Study, Survey Design Effects","lastPublishedDoi":"10.21203/rs.3.rs-6154489/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6154489/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study assesses the impact of the COVID-19 pandemic on the mental health of 16- to 29-year-olds in the United Kingdom, using longitudinal data from the UK Household Longitudinal Study (UKHLS) and its predecessor, covering the period from 2001 to 2023. The study identifies the causal effects of the lockdown (April 2020\u0026ndash;March 2021) and the post-lockdown period (April 2021\u0026ndash;March 2022) by estimating counterfactual mental health trajectories based on long-term trends. Unlike prior research, it accounts for potential reporting bias introduced by the UKHLS COVID-19 study. Mental ill-health among young adults had been rising for nearly two decades before the pandemic. During the lockdown period, the average General Health Questionnaire (GHQ-12) psychological distress score increased by 9% of its standard deviation, while the prevalence of clinically relevant psychological distress rose by 4.5 percentage points. This impact was temporary, with mental health levels returning to predicted trends by April 2021, suggesting no lasting 'scar' on average mental health. The recovery coincided with declining feelings of loneliness and increased life satisfaction. The study also identifies variations in the pandemic\u0026rsquo;s mental health effects by gender, household income, age, and ethnicity. Women and young adults in the top third of the household income distribution experienced a more pronounced increase in psychological distress during lockdown. However, there is no evidence that the under-30 age group suffered, on average, more severe mental health effects than the rest of the adult population under 60 during the lockdown period. The findings challenge prevalent narratives by demonstrating the relative resilience of young adults in the face of the pandemic.\u003c/p\u003e","manuscriptTitle":"Revisiting the Mental Health Impact of COVID-19 on Young Adults in the UK: Long-Term Trends, Temporary Setbacks, and Recovery","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-05 09:19:16","doi":"10.21203/rs.3.rs-6154489/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d31dc099-3c5d-45e8-9230-11f9acb7818f","owner":[],"postedDate":"March 5th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":45180989,"name":"Epidemiology"},{"id":45180990,"name":"Health Economics \u0026 Outcomes Research"}],"tags":[],"updatedAt":"2025-05-12T20:14:57+00:00","versionOfRecord":{"articleIdentity":"rs-6154489","link":"https://doi.org/10.1007/s11205-025-03616-8","journal":{"identity":"social-indicators-research","isVorOnly":false,"title":"Social Indicators Research"},"publishedOn":"2025-05-10 00:00:00","publishedOnDateReadable":"May 10th, 2025"},"versionCreatedAt":"2025-03-05 09:19:16","video":"","vorDoi":"10.1007/s11205-025-03616-8","vorDoiUrl":"https://doi.org/10.1007/s11205-025-03616-8","workflowStages":[]},"version":"v1","identity":"rs-6154489","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6154489","identity":"rs-6154489","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}