Assessing the Dimensionality of Dichotomous Data Using Modified Order Analysis

Abstract
Previous research has shown that order analysis and factor analysis typically yield different results in assessing the dimensionality of dichotomous data. This paper demonstrates that, through the inclusion of item proximity information, the dimensions identified by order analysis can be highly congruent with the results of factor analysis. A modified order analysis procedure is presented and compared to factor analysis using both simulated data of known dimensionality and real data.

This publication has 7 references indexed in Scilit: