Advancing and Evaluating the Applicability of Latent Response Models for Identifying Careless and Insufficient Effort Responding in Survey Data

• Applicant: Professor Dr. Gabriel Nagy, since 3/2024
• Subject Area: Personality Psychology, Clinical and Medical Psychology, Methodology
• Term: since 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 508980636

Careless and insufficient effort responding (C/IER) poses a well-known threat to the quality of questionnaire data. C/IER can manifest itself in a plethora of behavioral patterns, ranging from uniform random responding to marking distinct patterns, rendering its detection a challenging endeavor. Recently developed latent response mixture modeling (LR-C/IER) approaches pose sophisticated tools that leverage additional item-level information such as response times or item features for identifying and modeling C/IER behavior. By being agnostic towards the specific types of C/IER patterns in the data, allowing for detecting C/IER on the item-by-respondent level, taking the uncertainty in C/IER identification into account, and adjusting for respondents' attentiveness in the estimation of content trait levels, these approaches come with great potential for advancing the understanding and handling of C/IER. Nevertheless, having only recently been developed, the performance and recommended scope of application of LR-C/IER approaches are yet not well understood. The goals of this project are to further advance their applicability, investigate their robustness under various types of C/IER behavior, evaluate whether they pose a valid measure for real-life C/IER behavior, and derive guidelines for their application. For increasing the applicability of LR-C/IER approaches, we aim to implement maximum-likelihood-based estimation in widely employed statistical software and compare these implementations against previously employed Bayesian estimation. These evaluations serve to derive guidelines for model estimation and formulate general data requirements under which LR-C/IER approaches yield trustworthy results. For investigating the robustness of LR-C/IER approaches, we aim to evaluate their performance under types of C/IER behavior that violate model assumptions. For investigating validity, we aim to evaluate the capability of LR-C/IER approaches to detect real-file C/IER behavior in experimental empirical data gathered under standard instructions as well as under conditions aimed at either evoking or preventing C/IER, and investigate agreement with other established behavioral and self-report measures of C/IER.

• DFG Programme: Research Grants
• International Connection: USA
• Co-Investigator: Professorin Dr. Steffi Pohl
• Cooperation Partner: Professor Dr. Nathan Bowling
• Ehemalige Antragstellerin: Dr. Esther Ulitzsch, until 3/2024