Comparing Methods to Detect Insufficient-Effort Responses in Questionnaire Data

• In Event: Strategies to Improve the Quality of Data and Model Fit in Survey Research

Abstract

Quality of questionnaire data may suffer from participants’ insufficient effort responding (IER). We examined the sensitivity of various IER detection methods in detecting different IER patterns in a real Humorous Style Questionnaire (HSQ) data set. Results indicated the different IER detection methods identified between 1.03% and 25.12% of respondents as potentially IER respondents. Removal of IER responses could improve the performance of validity and reliability indicators. We strongly recommended the combination of different IER detection methods in empirical research. It gave a better coverage of different IER patterns by combining IRV with lz and MD. The adoption of both lz and IRV methods would improve both reliability and validity indicators.

Authors

• Dan Wang, The Chinese University of Hong Kong
  Presenting Author

• Kit-Tai Hau, Chinese University of Hong Kong
  Non-Presenting Author

• Leifeng Xiao, Chinese University of Hong Kong
  Non-Presenting Author

• Luyang Guo, Chinese University of Hong Kong
  Non-Presenting Author