A Strong Validation of the Crosswise Model Using Experimentally-Induced Cheating Behavior.

Area of Science:

• Psychological research methods
• Survey methodology
• Social psychology

Background:

• Socially desirable responding (SDR) poses a significant challenge in surveys, leading to inaccurate data.
• Indirect questioning techniques aim to mitigate SDR by reducing response bias.
• The Crosswise Model (CWM) is a promising indirect questioning method for sensitive topics.

Purpose of the Study:

• To validate the effectiveness of the Crosswise Model (CWM) in controlling for socially desirable responding.
• To compare the accuracy of CWM with direct questioning for sensitive attributes.
• To assess the prevalence of cheating in an online research paradigm.

Main Methods:

• An online cheating paradigm was developed to establish a known prevalence of a sensitive behavior (cheating).
• Participants attempted to identify target words from anagrams, with one unsolvable anagram to induce cheating for rewards.
• Direct questions and CWM indirect questions were used to assess cheating prevalence.

Main Results:

• A known cheating prevalence of 15.5% was established among 664 participants.
• Direct questioning underestimated cheating prevalence at 5.1%.
• The Crosswise Model (CWM) provided a more accurate estimate of 13.0% for cheating prevalence.

Conclusions:

• The Crosswise Model (CWM) effectively controls for socially desirable responding on sensitive questions.
• CWM yields significantly more accurate prevalence estimates of sensitive personal attributes than direct questions.
• The developed online cheating paradigm serves as a valid tool for testing indirect questioning techniques.