Kids These Days! A Cross-Temporal Meta-Analysis of Changes in Emotional and Behavioral Problems of Population-Based Samples Over the Past Decades

There is a general belief that mental health problems are increasing in younger generations, though little objective comparisons exist. The current meta-analysis examines changes in subclinical mean-level overall symptom burden and shifts in specific domains among population-based samples of youth (1-18 years) globally over the last four decades. We search systematically in PubMed, Web of Science, Scopus, EBSCO and Google Scholar with pre-registered criteria [removed for masked review]. Included studies (k = 175, N = 418,528) use the standardized and cross-culturally validated Child Behavior Checklist with parent-, teacher-, and self-report. A cross-temporal meta-analysis is conducted with data points between 1981 to 2019. Age, sex, response rate, and geographic continent, used as moderators. Meta-regression analyses, with year of data collection as a predictor, shows overall symptom burden did not change in the past 40 years except for some decreases in early childhood (parent report), middle childhood (self-report) and teenage boys (self-report). There is evidence of change in specific symptom domains. We find strong, multi-informant evidence for decreased externalizing problems (parent report and self-report in girls) including aggression in middle childhood over time (parent report and self-report), and internalizing problems (parent and teacher report of boys and girls) in middle childhood. Single-informant changes show varying patterns in subscales of internalizing problems with increases in anxiety/depression in teens (parent report) and somatic complaints in teenage boys (self-report) next to a decrease in withdrawn/depressed symptoms (self-report of children and youth overall and parent report in middle childhood). Further single-informant changes suggest an increase in social problems in European teens (self-report) and a decrease in rule-breaking behavior (parent report). Finally, contradicting results were found for attention problems (an increase based on parent report and a decrease in middle childhood boys based on teacher report) and thought problems (an increase based on parent report and decrease in teenage boys based on self-report). However, the finding regarding attention problems should be interpreted cautiously given declining reliability of the subscale over time and evidence suggesting potential measurement non-invariance. In sum, our results do not support concerns that young people today have more overall symptom burden than previous generations, rather, they show a different composition of symptom profiles. However, an important limitation of the study is that post COVID-19 changes were not studied.