Does intervening to enhance parental involvement in education lead to better academic results for children? An extended review

4,898 reports screened; 127 attempted evaluations of parental-involvement interventions

Children at or before primary school, internationally

A design-quality audit rather than a pooled estimate — the authors grade every evaluation they can find and report the distribution of quality, which is precisely the operation the rest of this literature omits.

"None of these studies was a large, robust evaluation." 121 of 127 reported research with serious limitations, and those 121 were "almost equally divided between those claiming success and those saying that the intervention had been ineffective or harmful." Of the six remaining, three offered positive outcomes, and those three were complex packages in which parental involvement was only one component — so not one clean, well-evaluated test of parental involvement alone produced a positive result.

Not the review's subject, but relevant: with a null-to-unknown causal effect from intervention studies, the large correlations in the observational parental-involvement literature are left with heredity and selection as the leading explanation.

Directly at odds with Nye et al. 2006's d=0.45. The disagreement is methodological — Nye pooled 19 short RCTs, See & Gorard judged the whole field including those trials to be too weak to pool.