Lifespan designs: age, cohort and repeated measurement
Distinguish age differences from within-person change using cross-sectional, longitudinal and sequential designs.
Question and prediction
Does a difference between today’s younger and older adults reveal how an individual changes with age?
Predict whether testing the same people again removes every alternative explanation.
Mechanism and evidence
Cross-sectional research compares age groups at one time. Longitudinal research follows people over time. A sequential design follows multiple cohorts, helping compare patterns across starting ages and historical periods. Each answers a different question. Age, cohort and period are related, so no simple observational design automatically separates every influence. Development can include gains, losses and changes in strategies rather than one uniform upward or downward line.
Repeated measurement estimates change among people observed more than once, but practice effects, selective attrition and changing measurement meaning remain. A speeded task may mix knowledge with sensory or motor demands. If a test becomes familiar, an improved score can reflect practice as well as development. A comparison needs a stable construct, not merely the same worksheet.
The NIH ABCD study provides an authentic example of following a youth cohort through development with repeated measures spanning health, environment and cognition. Its observational design does not randomly assign family experiences. The fictional table below is unrelated to ABCD: two age groups differ at one time, while the older group’s returning participants improve slightly. Both findings can coexist because they describe different comparisons.
Design a study around the mechanism you want to investigate. To examine practice, include appropriately comparable first-time testers; to examine historical differences, follow multiple cohorts; to examine attrition, compare who returns with who does not. These additions strengthen particular inferences without transforming age into a randomly assigned treatment.
Inspect the relationships
| Comparison | People | What changes? |
|---|
| Age 30 versus 60 in year 1 | Different groups | Age and cohort |
| Returning older group across years | Same people | Age, period and test experience |
| Returners versus nonreturners | Selected subsets | Retention composition |
Worked example
Cross-sectional gap=70−60=10. Matched returners’ change=62−61=1. Comparing 62 with the entire original older group’s 60 would mix within-person change with sample composition. Neither number alone identifies aging’s causal effect.
Your investigation
Draw a cohort-by-calendar-year grid. Label observations as between-person or within-person. Use the authentic ABCD design card to explain what repeated observation adds, then write a 200-word design separating practice from change without claiming random assignment of age.
Unfamiliar transfer
Two cohorts improve after taking the same test repeatedly. What additional comparison targets practice? Explain your reasoning before checking feedback.
Reading
NIH Brain Development Cohorts: study design. The records below are original fictional teaching cases, not research results from this source.
Fictional age comparison At year 1, adults aged 30 average 70 points and adults aged 60 average 60 on the same task. These are different people.
Fictional repeated sample The 60-year-old group’s 40 returning participants average 62 at year 3. Their own year-1 mean was 61; 20 other original participants did not return.
Authentic evidence and its limits
Official ABCD cohort design — NIH Brain Development Cohorts (Current study overview, checked October 2026). The NIH describes ABCD as a longitudinal study following youth from ages 9–10 into early adulthood, with repeated measures of health, environment, cognition and brain development.
Bounded paraphrase of the study sponsor’s design description, not an ABCD result. Repeated observation is not random assignment of life experiences. The numerical classroom table is unrelated to the cohort.
Read the attributed source. Original LearnAI paraphrase of the attributed source; no reproduced figure, participant data or verbatim passage. Original publication retains its rights.
