Response Speed and Response Consistency as Mutually Validating Indicators of Data Quality in Online Samples

We examine the appropriateness of response speed and response consistency as data quality indicators within online samples. Across several inventories, results show that response consistency decreases dramatically at response rates faster than 1 second per item. Our results suggest that careless responding may be fairly common in online samples and often functions to increase the expected correlation between items in a survey, which has implications for the likelihood of false positives and the analysis of factor structure. Given how careless responding can influence estimated associations between variables, we strongly recommend that researchers include response speed and consistency screens in their research and provide empirically informed cut points for data screens that should be useful across a wide range of instruments and settings.