M. Danish Shakeel · Educational Psychology Review 2022 · repeated cross-sectional cohort trend analysis · n=7,000,000 tests across 160 survey waves

A Half Century of Progress in US Student Achievement: Agency and Flynn Effects, Ethnic and SES Differences

Cited 25 times in the scientific literature.

Level 3 - non-randomized controlled study

Design analogy: repeated cross-sectional cohort trend analysis of national probability survey datasets.

OpenAlex W4214903318 · doi:10.1007/s10648-021-09657-y · record verified 2026-08-31

What was done

Analyzed approximately 7 million psychometrically linked standardized tests in mathematics and reading administered across 160 survey waves to national probability samples of US cohorts born between 1954 and 2007. Trends were evaluated by testing agency, subject, age at testing, birth cohort, race/ethnicity, and socioeconomic status (SES) quartile, measured as standard deviation (sd) change per decade.

What was found

Estimated change per decade varied across testing agencies (math: –0.10sd to 0.27sd; reading: –0.02sd to 0.12sd), with median overall gains higher in math (0.19sd) than reading (0.04sd). Cohorts born after 1990 experienced a deceleration in math progress and an acceleration in reading. Younger students made greater progress (math: 0.31sd, reading: 0.08sd) than middle-school students (math: 0.17sd, reading: 0.03sd) or end-of-schooling students (math: 0.06sd, reading: 0.02sd). Progress for young white students (math: 0.28sd, reading: 0.09sd) was lower than for Asian (math: 0.46sd, reading: 0.28sd), Black (math: 0.36sd, reading: 0.19sd), and Hispanic (math: 0.29sd, reading: 0.13sd) students, with ethnic differences narrowing as students aged. Young students in the lowest SES quartile gained more than those in the highest quartile (difference: 0.08sd in math, 0.15sd in reading), though this pattern reversed by the end of schooling.

Why it matters

This analysis synthesizes 50 years of national standardized testing to show that US academic achievement has broadly improved—particularly in math and at early ages—while revealing that early equity gains by race and SES largely dissipate by secondary school.

Limits

Results depend on psychometric equating across disparate testing programs and agencies, which showed considerable variation in estimated trajectories. The observational design cannot isolate the specific causal contributions of schooling, nutrition, healthcare, or environmental changes to the observed achievement trends.

Cited by