Autore
Contini, Dalit

Titolo
Cross-sectional learning assessments: comparability of regression coefficients and validity of difference-in-difference estimation to evaluate institutional effects.
Periodico
Università degli studi di Torino. Dip. Di Economia e Statistica Cognetti de Martiis. Working paper series
Anno: 2014 - Volume: 10 - Fascicolo: 31 - Pagina iniziale: 1 - Pagina finale: 26

The way achievement inequalities among socio-demographic groups develop throughout childhood in different institutional contexts is a matter of great interest from a policy perspective. Yet, international l earning assessments, like many national studies, are cross-sectional and non-vertically-equated (i.e. achievement i s not measured on a unique scale in different surveys at different age or grades). Against this background, the aim of this paper is twofold. First, I show that the comparison of regression coefficients with non-vertically-equated achievement scores as dependent variables does not convey much information on the development of disparities related to specific individual characteristic, even when applied to standardized scores. On these grounds, I question the validity of difference in difference estimation whose fundamental element is the comparison of regression coefficients between two non equated learning assessments administered at different grades for the evaluation of the impact of institutional features on achievement inequalities related to ascribed individual characteristics like gender or socioeconomic background.



Testo completo: http://www.est.unito.it/do/home.pl/Download?doc=/allegati/wp2014dip/wp_31_2014.pdf

Esportazione dati in Refworks (solo per utenti abilitati)