Revisiting the integration hypothesis again: High heterogeneity complicates the interpretation of cross-sectional evidence
After our meta-analyses weakened the claim that integration produces the best adaptation, critics answered that even a small effect matters across millions of migrants. This rejoinder tests that defense and finds it wanting: the true effects of integration are so heterogeneous that they are estimated to be negative almost 30% of the time.
The paper, in plain language
The 'small but robust' defense
When Kinga Bierwiaczonek and I showed meta-analytically that acculturation strategies relate only weakly to adaptation, defenders of the integration hypothesis replied that small effects still matter at population scale. With Mike Cheung — the methodologist behind the meta-analytic structural equation modelling tools we used — we examined the premise hiding in that defense: that the small effect is at least consistent.
Unpredictable is the finding
It is not consistent. Modelling the distribution of true effects, integration and heritage-culture orientation are estimated to relate negatively to adaptation in nearly three cases out of ten; only mainstream-culture orientation behaves homogeneously. As we put it in the paper: for some migrant groups integration will accompany positive outcomes, for others negative ones — with no evidence that either outcome is caused by integration.
Why it matters
A treatment that helps some patients and harms others has no business being prescribed to everyone, and the same logic applies to integration policy built on cross-sectional correlations. This exchange — critique, defense, rejoinder, with John Berry himself among the interlocutors — is how the field is renegotiating its most cited claim, and it connects directly to my argument that acculturation research faces a causality crisis. This summary rests on the paper's abstract and our own research summaries, as the article's full text sits behind access barriers even we could not cleanly link.