Methodenzentrum

08.11.2022

New Publication: Automated Bot Detection Using Bayesian Latent Class Models in Online Surveys. Roman, Z. J., Brandt, H., & Miller, J. M. (2022)

In this article we propose a Bayesian latent class structural equation model that is able to detect bots on surveys. One of the main features is that the detection is based only on the response pattern that both bots and humans provide. This new approach is evaluated with a simulation study and an empirical data set where the true bots could be identified via duplicated geolocations and/or IP addresses. The results indicate a high accuracy for bot detection which makes this approach a useful tool in applied sciences.

 

Roman, Z. J., Brandt, H., & Miller, J. M. (2022). Automated Bot Detection Using Bayesian Latent Class Models in Online Surveys.

Zurück