The Data Cleaner uses statistical heuristics to identify low-quality respondents.
1. Speeders
Respondents who complete the survey too quickly. We calculate the median Length of Interview (LOI) and flag anyone below a certain threshold (e.g., 40% of the median).
2. Straight-Liners
Respondents who give the same answer to every question in a grid (e.g., all "5"s). We calculate the standard deviation of their responses for a specific grid; if SD = 0, they are flagged.
3. Gibberish Detection
We check open-ended responses for patterns like repeated characters (e.g., "asdfasdf") or lack of vowels/consonants that indicate non-human or low-effort input.