The new addition to the ReDem in-survey quality tool compares answer distributions before and after quality controls are applied, allowing researchers to identify questions where removing low-quality respondents changed the distribution of responses.

Rep Data acquired ReDem earlier this year. ReDem’s in-survey response evaluation is designed to detect and block bots and AI-assisted survey takers.

As part of its process to develop the new feature, Rep Data analysed around 12,700 survey questions and found that one in 16 questions changed its leading response. Individual answer options also shifted by more than five percentage points in some cases, with the largest observed change reaching 11.7 percentage points.

Sebastian Berger, head of science for ReDem by Rep Data, said: "Researchers already have processes for identifying and removing low-quality interviews.

"What’s often missing is an easy way to understand whether those quality decisions affected the findings. By comparing results before and after cleaning, researchers can quickly see where quality controls had little impact and where they meaningfully changed the outcome."