Reconsidering Focus Groups and Rasch Model Item Response Theory in Instrument Development

Abstract

In this rejoinder, we focus on five important themes derived from the valuable comments of Harvey; Wang, Hogge, and Sahai; and Whittaker and Worthington. Specifically, we reflect on (a) methodological and conceptual issues associated with the use of focus groups in developing an initial item pool, (b) the debate about the use of Rasch versus other item response theory (IRT) models, (c) response scale functioning and the use of item parcels, (d) bandwidth and scale sensitivity, and (e) multicultural considerations. We reconsider the strengths and limitations of the approaches that we have endorsed in light of the comments of our colleagues. We conclude with the hope that dialogue on the use of focus groups and IRT in measurement development and on multicultural assessment continues.

Coercive Control in the Courtroom: the Legal Abuse Scale (LAS)

“…See Table 2 for item z-scores as well as infit and outfit statistics. The item separation index (4.92) and reliability (0.96) indicated that the item difficulties can be rank ordered on the latent trait (Meyer, 2014). The person separation index (1.51) and reliability (0.69) supported that the items could adequately discriminate between those with high and low levels of legal abuse (Fisher, 1992).…”

“…Because extant tested measures for legal abuse are limited, EFA is an appropriate analytic approach for the present study as it determines construct validity during initial scale development (Worthington & Whittaker, 2006). Quantitative analyses were conducted primarily in STATA 16.0 (StataCorp, 2019) and jMetrik (Meyer, 2014).…”

Recent Developments in the Assessment of Nutrition Knowledge in Athletes

“…Focus groups were used by Trakman and colleagues [19] to assess the clarity of the NSKQ after the item pool was generated, whereas the authors of the PEAKS-NQ used focus groups with SN to inform the initial item pool [31]. Focus groups can be useful in generating a representative item pool by giving participants an opportunity to stimulate each other's thinking, providing more diverse perspectives that uncover researchers' "blind spots" and can reduce researcher bias [43][44][45]. This method remains underutilised in this research area.…”