Social Studies Instruction and Reading Comprehension: Evidence from the Early Childhood Longitudinal Study

Analytic sample of 6,731-6,829 students drawn from ECLS-K:2011, which sampled over 18,000 kindergarteners

US students followed from kindergarten (2010-11) through fifth grade

Observational analysis of the Early Childhood Longitudinal Study, regressing fifth-grade reading achievement on reported instructional minutes per subject, with student- and school-level controls. Not a study of source work at all - the exposure is TIME SPENT ON SOCIAL STUDIES, not document-based or inquiry instruction - but it is the single most cited empirical support for the claim that history/social studies instruction builds reading comprehension, so it needs a record. The authors are candid about the identification problem: instructional time "is not randomly assigned and may correlate with other factors for which we are unable to control - such as teacher quality", so the association "could imply a causal effect or, at the very least, suggest how better teachers" allocate time. Under this archive's premise the confounding is worse than that concession admits: schools and teachers that protect social-studies time are selected on intake and on parental characteristics, and knowledge-rich home environments are transmitted genetically as well as environmentally.

Thirty additional daily minutes of social studies was associated with fifth-grade reading scores about 15% of a standard deviation higher, and it was the only subject with a clear, positive, statistically significant association; extra ELA time showed roughly 3% of an SD and was not distinguishable from zero. Associations were larger for lower-SES students (17-21% of an SD in the bottom three income quartiles). The result is genuinely interesting as a hypothesis about knowledge-building and reading, and it is correlational.

High. Instructional time is chosen by schools and teachers, not assigned; the same schools and families that produce more social studies time produce children who read better for reasons the covariates cannot absorb. This is the archive's canonical case where a correlation is not evidence of a teachable lever.

Unreplicated as a causal claim, and not testable as one from these data. The closest experimental analogue is the PACT programme, where reorganising middle-school history around text raised content knowledge and left Gates-MacGinitie reading comprehension untouched.