To investigate the effects of the task of initialing a contract on consumer behaviour, we set out to explore whether the addition of a secondary task, in general, would impact recall of the information in contract clauses. However, what seemed like exciting, statistically significant results could not be trusted due to the influence of bots. Unfortunately, this is in line with reported shifting quality of data from online labor markets.

Nevertheless, the things that threatened our results (the bots) also showed us how to outsmart them. We explain why, and what can be done