Open-ended research is essential to building solid hypotheses in the social sciences — without it, even the best-planned analyses can fail.