An item level evaluation of the Marlowe-Crowne Social Desirability Scale using item response theory on Icelandic Internet panel data and cognitive interviews

•Items from the MCSDS were analyzed with IRT and cognitive interviews.•Ten items from the MCSDS were shown to have poor psychometric properties.•A psychometrically sound short form of the MCSDS was developed.

The Marlowe-Crowne Social Desirability Scale (MCSDS) is commonly used to validate other self-report measures within social and health research. Concerns over the scale's length (33 items) have repeatedly been raised. Nevertheless, prior efforts to develop psychometrically sound short forms of the MCSDS have not led to consistent findings. The purpose of this study was to develop a short form of the MCSDS, in accordance with guidelines for best practices in short form and scale development. Information on item properties, obtained with item response theory (IRT) and cognitive interviews (CogI), were used to eliminate items with poor properties and select items for a short form to be administered via the Internet. The IRT analyses were based on responses from 536 Internet panel members and the CogI sample consisted of 40 interviewees. Ten items were dropped due to poor psychometric properties and out of the 23 remaining items a ten item short form was developed.