X-Culture collects immense amounts of data. Over 2,000 variables. Multi-Level. Multi-Source. Multi-Method. Longitudinal.
We have weekly surveys of all X-Culture participants.
However, much of it is self-report questionnaires. As we learned the hard way, self-report questionnaires are tricky. It’s not that people intentionally lie. But the threat to validity of the data posed by the subjective interpretation of the questions, anchoring, acquiescence bias, extreme response bias, comparison bias, socially-describable answers, and good old response fatigue are HUGE.
Bad data in, misleading results out. Garbage in, garbage out.
In this blog, I will tell you how we detect random (or systemically bias) response, tricks we use to minimize the problem, as well as share some funny (and some troubling) stories about self-report data collection.