Effects of grant aid on persistence and completion are heterogeneous across studies, justifying random-effects meta-analysis¶

Claim · All claims
Evidence · 1 study · quant-synthesis r3 · q3
Currency · newest study 2018 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? For on-time degree completion, I2 across 20 studies is 85.136, indicating less than 15 percent of variation is random error. → Nguyen 2018

Evidence¶

Nguyen 2018¶

Nguyen, T.D., Kramer, J.W., & Evans, B.J. (2018). The Effects of Grant Aid on Student Persistence and Degree Attainment: A Systematic Review and Meta-Analysis of the Causal Evidence (CEPA Working Paper No. 18-04). Stanford Center for Education Policy Analysis. http://cepa.stanford.edu/wp18-04

q3 · i? · quant-synthesis · r3

Heterogeneity statistics reported in the Results for the on-time completion outcome of the random-effects meta-analysis; the printed I2 of 85.136 across 20 studies indicates most observed variance reflects true heterogeneity rather than random error. Similar evidence was found for all four outcomes.

"in terms of on-time degree completion the percentage of observed variance across 20 studies that reflects true heterogeneity in effect sizes (I2) is 85.136, indicating that less than 15 percent of the total variation can be attributed to random error."

Discussion¶

Related Claims¶

• Grant aid increases degree completion, raising on-time completion by 2.4 and delayed completion by 2.5 percentage points in U.S. causal studies — related
• Grant aid programs bundling additional academic or social supports show larger persistence and completion effects than aid alone — related
• Grant aid increases postsecondary persistence, raising within-year persistence by 3.2 and year-to-year persistence by 1.8 percentage points in U.S. causal studies — related
• Need-based grant aid improves persistence and completion while merit-only aid shows no detectable effect on these outcomes — related
• An additional $1,000 of grant aid improves year-to-year persistence by a statistically significant 1.2 percentage points, with smaller statistically insignificant effects on degree completion — related
• Funnel-plot analysis finds no evidence of missing small-effect studies, though asymmetry from few negative estimates suggests possible small bias against null-negative results — related
• Effect sizes across shared text reading studies show substantial true heterogeneity (I2 = 80.09) — related
• Effect-size heterogeneity across Covid-19 studies is essentially total (I2 = 100%), justifying a random-effects model — a narrower finding that bears on this claim