Answering Big Questions with Small Samples
Another Friday, another episode of the Deep-Thinking-project.
Last weekend I was interviewed by a Dutch Newspaper (Financieel Dagblad) about my work as a statistical consultant and how I deal with limited data sets as opposed to Big Data. My answer was that if you suffer from a limited dataset, for whatever reason, it becomes more and more important to ask better (more precise) questions.
I received dozens of emails from people who agreed that asking questions is indeed very important and often forgotten. As soon as data comes in we immediately start searching for patterns. Instead, I propose that we should spend much more of our research time on discussing which questions we are actually interested in.
I will immediately put my own suggestion into practice. Next week, I start working on a revision for one of my papers and instead of directly starting to implement all the feedback from the reviewers, I will first ‘sit and think’ about which question I actually wanted to address in this paper. Only if reviewer comments contribute to answering the original question, I will implement these in a new version of our manuscript. If not, I will gently inform the reviewer that his/her suggestion is ignored.
Any suggestions how to phrase this in a polite way are very much appreciated!