Analyzing a German-language Expanded Form of the PhoPhiKat-45: Psychometric Properties, Factorial Structure, Measurement Invariance with the Likert-Version, and Self-Peer Convergence

Abstract

Personality assessment typically relies on self-report questionnaires utilizing Likert-type scales. Recently, the Expanded format has been proposed as alternative, but research on the consequences of adapting Likert-like responses to Expanded items is sparse. We adapt a multidimensional measure into the Expanded format: the PhoPhiKat-45. This is the standard questionnaire to assess gelotophobia (fear of being laughed at), gelotophilia (joy in being laughed at), and katagelasticism (joy in laughing at others). We test the reliability, item/scale parameters, and factorial structure across the Expanded and Likert formats in three independently collected samples (Ns = 323/261/460). While the psychometric properties are satisfying, elevated item- and mean-scores in our experimental Expanded version do not support full measurement invariance with the Likert version—and, thus, it does not permit the application of established cutoff scores for gelotophobia. The convergence of self-peer ratings supports the validity of an Expanded version. Overall, an Expanded form of the PhoPhiKat-45 did not outperform the standard Likert version. We discuss potential trait- and method-related causes for the measurement invariance and consequences for the usage of cutoff scores in the assessment of gelotophobia as well as interpreting the findings in light of continued research on the Expanded response format.

Acknowledgments

We are grateful to Katja Glathe and Jasmin Leuschke for their help with the data collection, Prof. Dr. Holger Kersten and Dr. Tracey Platt for proofreading the preliminary English-language Expanded format items, and to Rebekka Sendatzki and Valerie Weigel for their help in preparing the tables. This study was not pre-registered.

Data availability statement

Data, materials, and syntaxes are available at the Open Science Framework under osf.io/9pw4d/

Open Scholarship

This article has earned the Center for Open science badges for Open Data and Open Materials. The data and materials are openly accessible at https://osf.io/9pw4d/ and https://osf.io/9pw4d/.

Notes

1 We also subsume the Likert-type response format under this term throughout the manuscript.

2 A similar pattern exists when comparing the data of Sample 1/2 (Expanded version) with Ruch and Proyer’s (2009) validation data of the Likert version (g[Pho] = 0.33/0.41, g[Phi] = 0.25/0.17, g[Kat] = 0.29/0.29). One reviewer suggested that we examine whether missing equidistance of the response options in our new Expanded version might explain the elevated scores. Using the threshold values τ estimated in the CFAs (Andrich, 1978; outputs available in the OSF), we found symmetric distances between the response categories (Sample 1/2: Δ[τ1-τ2] = 1.17/1.17, Δ[τ2-τ3] = 1.19/1.16). Thus, response categories are considered equidistant in the proposed Expanded version.