I see this as a common problem in #evaluation too - premature focus on statistical significance of differences in means without checking distributions (bi-modal? skewed?) or bivariate or multivariate patterns. Something I used to emphasise when teaching statistics to students.

Exactly! Thanks for sharing this helpful paper.

Thank you @guerzovich.bsky.social for this great bringing together of #rubrics and #theoryofchange with an important addition - #typologies! It all makes such sense now - and shows the value of simple cataloguing as an initial task in #qualitativedatanalysis. #evaluation #impactevaluation

Evidence of effectiveness in relation to stated policy goal is clearly not sufficient.

Agree it's a poor label, since all #evaluation is based on some version of #theory. It was originally used for evaluations that at least claimed to be framed around an explicit #programtheory or #theoryofchange. But it's now being used unhelpfully for non-experimental impact evaluation designs.

Congratulations to @sabincenter.bsky.social for creating three online tools to track decisions and actions of the new USA government - an important contribution to #accountability.