Bots Impact On Online Surveys Within Special Education Research

Abstract

Researchers use online surveys to collect data from diverse populations (Phillips, 2017; Simone, 2019a). With the use of online surveys comes the potential for Artificial Intelligence Bots (AI Bots) to enter a survey and pose as valid participant. In this study we asked participants to inform us about their online research and if bots impacted their research in any way. To be eligible for the survey participants had to conduct an online survey in the field of special education research from 2010 to present. The participants could report up to ten surveys they have conducted since 2010 and if bots impacted their online survey. Out of the 88 surveys reported, 50 surveys (56.8%) had known or suspected bots. Of the participants who reported having known or suspected bots the mean percentage of bot responses in their data set was between 21%-40% of survey responses. When participants had known or suspected bots in their survey they reported having an increased amount of time during data cleaning. Having an increased number of bots in your survey can lead to data results being skewed (Storozuk et al., 2020). Currently, there is no research on how published online research results might be skewed due to bots posing as participants.