# Examining a Dutch short form of The Balanced Inventory of Desirable Responding Version 6: comparing polytomous and dichotomous scoring methods in a multidimensional framework

## TL;DR

This paper creates and validates a 20-item short form of a Dutch social desirability measure, comparing different scoring methods and evaluating its psychometric properties.

## Contribution

The paper introduces a validated short form of the Dutch BIDR and compares dichotomous and polytomous scoring methods in a multidimensional framework.

## Key findings

- A 20-item short form (BIDR-D20) was developed with equal or better psychometric qualities than the original 40-item version.
- Dichotomous scoring outperformed polytomous scoring in model fit and internal consistency.
- The short form correlated with self-reported aggression but failed to detect response bias in the sample.

## Abstract

The 40-item Balanced Inventory of Desirable Responding Version 6 (BIDR) is a widely used tool to measure two components of social desirability: Self-Deceptive Enhancement (SDE) and Impression Management (IM). In three studies, we aimed to create and validate a short form of the Dutch language version of the 40-item BIDR.

In Study 1 (general population sample N = 577), item properties were examined using (Multidimensional) Item Response Theory (IRT) for both dichotomous and polytomous scoring methods to create a short form. In Study 2 (general population sample N = 719), IRT analyses of Study 1 were replicated, and the nomological network of the short form was examined by investigating its relation with the Big Five personality traits and deviant traits and thoughts. Study 3 (men from the general population N = 100) investigated whether SDE and IM could detect response bias in self-reported aggression. All samples consisted of individuals volunteering to participate in scientific research (recruited in various ways) in a low-stake condition.

This yielded a short form containing 10 SDE and 10 IM dichotomously scored items (BIDR-D20). While results indicated a loss of information compared to the original version, the overall psychometric qualities were equal to or sometimes better compared to the BIDR (Version 6). Across studies, dichotomous scoring was generally better than polytomous scoring in terms of model fit, estimated IRT parameters, and internal consistency. Both forms correlated with self-reported aggression, but SDE and IM failed to detect response bias in the current sample.

The BIDR-D20 could be a worthy replacement for the 40-item BIDR (Version 6), with the same properties and less time-consuming. However, more research is needed to establish the short measure’s predictive validity as a response bias.