A formal modeling approach towards assessing socially desirable responding with the overclaiming technique

• Applicant: Professor Dr. Morten Moshagen
• Subject Area: Personality Psychology, Clinical and Medical Psychology, Methodology
• Term: from 2014 to 2018
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 260876406

Although interviews and questionnaires are widely used in behavioral research, the validity of self-reports of sensitive attitudes, attributes, and behaviors suffers from the tendency of individuals to provide overly positive self-descriptions. One approach to deal with this issue is to assess (and ultimately statistically control for) individuals´ tendencies to provide socially desirable responses. The overclaiming technique aims to measure such tendencies by considering the discrepancy between a self-report and an external criterion. To this end, participants receive a list made up of both factually existent and invented words (allegedly referring to general knowledge) and are asked to state whether they are familiar with each of the items. Overclaiming is defined as the tendency to claim knowledge about factually non-existent items, and, consequently, can be used as a criterion discrepancy measure of self-presentation tendencies. The immediate advantage over traditional approaches (such as widely used impression management scales) lies in the fact that systematic differences between self-report and criterion are likely due to socially desirable responding, rather than due to possessing actual personality traits. Despite this conceptual advantage, the usual operationalization of overclaiming via an index based on signal detection theory suffers from the confound of response tendencies and memory processes. In addition, it is unclear how to interpret the entire range of this index. The planned project attempts to overcome this limitations by developing and validating an improved measurement model of overclaiming, rooted in the multinomial processing tree modeling framework. The proposed model decomposes memory processes from response tendencies and additionally allows for an unambiguous interpretation of the entire parameter space. The first part of the project covers the systematic, mainly experimental, validation of the parameters of the proposed measurement model to ensure their interpretation as psychological process components. The second part comprises a comprehensive set of studies that (a) attempt to demonstrate the utility of the overclaiming construct in measuring and controlling response distortions and (b) comparatively evaluate the proposed measurement model against the overclaiming index derived from signal detection theory. Provided that the proposed measurement model proves useful in reliably assessing the tendency to provide socially desirable responses, this can be seen as substantial improvement over existing approaches directed towards this issue.

• DFG Programme: Research Grants