Applications of the Partial Credit Model (PCM) Accounting for Extreme Response Styles (ERS) with a Constrained Weight Parameter across Multiple Scales (ERS-PCM-S)

Files
KIM-DISSERTATION-2024.pdf (11.63 MB)
Author(s)
Kim, Eunbee
Advisor(s)
Editor(s)
Associated Organization(s)
Organizational Unit
Organizational Unit
Series
Collections
Theses and Dissertations
Supplementary to:
Permanent Link
https://hdl.handle.net/1853/78629
Abstract
It is worth noting that the structural dependence between extreme response style (ERS) and the measured trait is intrinsic to measurements using rating scales. Respondents with high or low trait levels tend to endorse more extreme options compared to those with middle trait levels. Polytomous models with thresholds, such as the partial credit model (PCM), have been extended to account for ERS tendency (ERS-PCM) by incorporating weight parameters multiplied by thresholds. Given that the weight parameter estimates in the ERS-PCM are affected by structural dependence, this study proposes the application of the ERS-PCM with the ERS tendency constrained to be constant across scales (ERS-PCM-C). In Study 1, the ERS-PCM-C was applied to measure Big Five personality traits. The findings indicate that the modified ERS-PCM-C reduces structural dependence in weight parameter estimates and produces different adjustments at the trait level, leading to a superior model fit compared to the ERS-PCM. Study 2 examines parameter recovery for the ERS-PCM-C under various simulated conditions, varying in the number of scales, mean trait levels within scales, and the relationship between ERS and the trait. The results demonstrate that the ERS-PCM-C accurately recovers weight parameter estimates and true correlations between ERS and the trait with reduced bias compared to the ERS-PCM. In the ERS-PCM, there is estimation bias in weight parameters. Furthermore, the findings illustrate that the ERS-PCM-C enhances trait parameter estimation, particularly for respondents with high and low weight parameters, resulting in different adjustments at the trait level compared to the ERS-PCM. The recommendation for models with constrained weight parameters underscores potential issues stemming from structural dependence between ERS and the trait, thereby enhancing the applicability and generalizability of IRT models with ERS extensions.
Sponsor
Date
2024-10-24
Extent
Resource Type
Text
Resource Subtype
Dissertation
Rights Statement
Rights URI