An examination of the internal consistency and structure of the Statistical Anxiety Rating Scale (STARS).

Abstract

The purpose of the current study was to examine the internal consistency and structure of the English version of the Statistical Anxiety Rating Scale (STARS). Participants were 202 (79% females) psychology undergraduates was recruited from James Cook University's Singapore (71%) and Australia (29%) campuses. Acceptable internal consistency reliabilities, ranging from .81 to .94, were found in this sample. Approximate fit indices suggest that a correlated six first-order factor model best describes the data in contrast to theoretical considerations suggesting that a six factor model with two correlated superordinate factors (i.e., statistics anxiety and attitudes toward statistics) best describes the data. Researchers are recommended to use part one of the STARS to assess statistics anxiety and part two to assess attitudes toward statistics.