Measuring Statistics Attitudes: Structure of the Survey of Attitudes Toward Statistics (Sats-36)

Abstract

Although a number of instruments for assessing attitudes toward statistics have been developed, several questions with regard to the structure and item functioning remain unresolved. In this study, the structure of the Survey of Attitudes Toward Statistics (SATS-36), a widely used questionnaire to measure six aspects of students’ attitudes toward statistics, is investigated. This study addresses the previously unexplored issue of individual item functioning. Based on confirmatory factor analysis using individual items, the results suggest that the SATS-36 can be improved by removing some poorly functioning items and that depending on the goals of a specific study either six subscales could be used or three of them (Affect, Cognitive Competence, and Difficulty) can be combined into one subscale without losing much information.
 First published May 2011 at Statistics Education Research Journal: Archives

Challenges Associated With Measuring Attitudes Using the Sats Family of Instruments

“…The final item-scale mapping is used for creating scores for each construct and ensures comparability across studies that use the same instrument. CFA results showed a four-factor structure as expected (Schau et al, 1995), but item parceling was used, which is no longer strongly recommended (VanHoof et al, 2011). Using an EFA, Cashin and Elmore (2005) found only two factors on the post-course SATS-28: Difficulty, all but one Cognitive Competence item, and most Affect items formed one factor while the Value items and the three remaining items formed the other.…”

“…Nolan et al noted the strength of NGT to develop the items relative to the process used by other instruments and an articulated educational theory for the constructs. Nolan et al also noted numerous studies about the factor structure of the SATS-28 and SATS-36 are consistent with four and six factor solutions, respectively, though the studies were not unanimous in this finding (e.g., Vanhoof et al, 2011) and were the subject of debate within the community at the time (cf. Schau & Millar, 2011).…”

“…Most studies that examined the factor structure of the SATS-28 and SATS-36 instruments included only undergraduate students and generally identified factor structures consistent with the proposed four and six factor solutions, respectively (e.g., Dauphinee et al, 1997;Hilton et al, 2004;Vanhoof et al, 2011). Few studies included graduate students in analyses of the structure of SATS instruments (e.g., Coetzee & Van der Merwe, 2010): one study with a sample of 342 students had approximately 80% graduate students, and an EFA yielded a two-factor solution for the SATS-28 (Cashin & Elmore, 2005).…”

“…In more recent literature using the SATS-36, studies have shown that both four-factor (Cashin & Elmore, 2005;Vanhoof et al, 2011) and six-factor (Carillo et al, 2016;Persson et al, 2019;Vanhoof et al, 2011;Xu & Schau, 2019) structures are supported. Studies showing support for a six-factor structure, however, tend to find better model fit for data from the post-course survey than from the pre-course survey, and researchers tend to identify several items that do not fit well (sometimes proposing them for deletion).…”

Attitudes toward Statistics and Statistical Literacy of Public Senior High School Students

“…Previous studies focused on determining attitudes toward statistics (Aslemand, 2018;Schau et al, 2003;Vanhoof et al, 2011;Khavenson et al, 2012) using various scale measures, but most utilized undergraduate students taking an introductory statistics course. This study adds up to the current body of knowledge of associating attitudes and statistical literacy utilizing senior high school students where a dearth in the literature was found.…”