When Does Inequality Grow? School Years, Summers, and Achievement Gaps

over 2.5 million students in about 21,000 public schools, three overlapping cohorts covering nine school years and six summers

US kindergarten through eighth graders taking NWEA MAP Growth, nationally weighted

NWEA Collaborative for Student Growth research brief; the journal version appeared in Educational Researcher. Full text of the brief read; the journal article was not. Seasonal decomposition on NWEA MAP Growth, an adaptive assessment on an equal-interval vertical scale, so the two artifacts that undermine the classic summer literature - age-spreading score scales and change of test form across the summer - do not apply. Descriptive rather than causal, hence the outcome classes below are measurements of when gaps move, not effects of anything.

THIS IS THE RESULT THAT KILLS THE SUMMER-GAP STORY, and it does so on the largest properly scaled dataset available. 'The academic achievement gap between Black and White students WIDENS WHEN SCHOOL IS IN SESSION and SHRINKS DURING SUMMERS.' The Black-White gap runs -0.54 SD in maths and -0.41 SD in reading at kindergarten entry and grows to -0.70 SD (30% larger) and -0.57 SD (39% larger) by the end of eighth grade - and that growth happens in school, because 'in summers, in contrast, White students showed larger summer drops than Black students, decreasing gaps'. The Asian-White gap tells the opposite seasonal story: 0.10 SD in maths and 0.09 in reading at entry growing to 0.45 SD and 0.21 SD by grade 8 - growth of 339% and 145% - and it happens 'primarily during summers'. Hispanic-White gaps declined slightly and showed no distinct seasonal pattern. WHAT SURVIVES FOR EVERYONE: 'all groups of students showed positive learning rates in school, and negative learning rates, or losses, during the summers.' Summer loss is real and universal; summer as the engine of the Black-White gap is not.

Not addressed and not addressable in this design, but the pattern is worth stating without a causal claim. Gaps that are largely present at kindergarten entry and then grow modestly are consistent both with early environmental investment and with heritable transmission; a seasonal decomposition distinguishes WHEN a gap changes, not WHY. The finding that gaps grow during school rather than outside it is a genuine constraint on the 'schools as equalisers' framing, and it does not by itself implicate schools as causes either.

Converges with von Hippel and Hamrock's conclusion that the evidence is 'inconsistent regarding whether gaps grow faster during school or during summer', with Quinn and colleagues' finding that seasonal inequality patterns in ECLS-K:2011 do not reproduce those in ECLS-K:1998, and with Atteberry and McEachin's estimate that demographics explain only 4% of summer-loss variance. Four independent analyses on three data sources. It directly contradicts the socio-economic half of Cooper's 1996 meta-analysis and the Baltimore Beginning School Study result that founded it.