Further validating the VAIAK: Defining a psychometric model, configural measurement invariance, reliability, and practical guidelines.

Abstract

Art interest and art knowledge are central constructs in empirical aesthetics and specifically the psychology of art. Therefore, valid measurement of these constructs is essential. To this aim the Vienna Art Interest Art Knowledge (VAIAK;Specker, Forster, Brinkmann, Boddy, Pelowski, et al., 2020) questionnaire has recently been introduced. Here, I present further validation evidence for this questionnaire in several ways. First, I define a psychometric model for the VAIAK which I consequently test by way of confirmatory factor analysis (CFA), which indicated good model fit and thus provided evidence for my theoretical model. To substantiate the CFA results, I also use exploratory factor analysis (EFA) models. Second, I address measurement invariance-specifically configural measurement invariance-between a group of art experts and non experts. In this case, too, the multigroup CFA indicated good fit and thus provides evidence for configural measurement invariance. Third, I address the issue of reliability from a theoretical standpoint and an empirical standpoint, specifically providing evidence that the VAIAK shows good test-retest reliability. Finally, I address some practical issues in using the VAIAK ("How long does it take?", "Can I shorten the VAIAK?") that researchers might encounter when using the VAIAK in their own work. In sum, the article presents a crucial next step in the validation process of the VAIAK.

Updating the Aesthetic Fluency Scale: Revised long and short forms for research in the psychology of the arts

“…Based on our test information curve, people with low knowledge (little experience with art; 2 SD below average) and very high knowledge (e.g., art experts; 4 SD above average) are measured relatively less reliably. As a result, this scale should not be used to define art expertise or to create extreme groups (i.e., selecting low scorers as “art novices” and high scorers as “art experts”) [ 10 ]. Finally, omega hierarchical ( ω h ), as assessed in this paper, is recommended for assessing internal consistency.…”

“…Another approach is to test people’s knowledge on questions about art that have correct answers. The VAIAK, for example, contains a section that assesses objective art knowledge [ 9 , 10 ]. Finally, a third approach is to measure what Specker et al [ 9 ] call subjective knowledge —people’s self-reported knowledge of the arts.…”

The next step for the VAIAK: An item-focused analysis.

“…Because prior work on the VAIAK has shown that items load either on the art interest or art knowledge factor and that the two factors are correlated (Specker, 2021), it is necessary to account for this covariance between the subscales when estimating the item parameters. To account for this in the IRT models, we simultaneously estimated the parameters for both the art interest and art knowledge subscales and allowed these latent factors to covary.…”

“…This could be problematic—to assess the scale properly, we need a sufficient sample size for both groups. For the purpose of this paper, we focus on formal education to demarcate expertise: if people indicate they had formal art education (with or without a degree) they are considered “art experts;” people indicating they did not have formal education are considered novices (see also footnote 2 and Specker, 2021).…”

Updating the Aesthetic Fluency Scale: Revised Long and Short Forms for Lab and Field Research in the Arts

“…Another approach is to test people's knowledge on questions about art that have correct answers. The VAIAK, for example, contains a section that assesses objective art knowledge (Specker, 2021;Specker et al, 2020). Finally, a third approach is to measure what Specker et al (2020) call subjective knowledge-people's self-reported knowledge of the arts.…”