Attached are some passages from the book, Cribsheet by Emily Oster, an economist who wrote a data-driven guide to parenting. I put together some interesting passages from the introduction and from one of the chapters that does a nice job contextualizing the concepts of data driven decision making, what a good study is, the limits of those studies, and the ultimate uncertainty of all the knowledge produced using data.
Meaningful connections to constructing knowledge and data collection in the human sciences (particularly economics), natural sciences, and cognitive biases. Also deals well with problems of sorting out the differences between correlation and causation.
Generally great book for parenting, not just for its TOK connections.
How can you identify a good study? This is a hard question. Some things you can see directly. Certain approaches are better than others – randomized trials, for example, are usually more compelling than other designs. Large studies tend, on average, to be better. More studies confirming the same thing tends to increase confidence, although not always – sometimes they all have the same biases in their results