In our day-to-day lives, most of us care deeply about a deceptively simple bottom-line question: Does something (e.g., a strategy, tool, or product) work? We use a certain toothpaste – does it reduce cavities more so than some other toothpaste? We use a certain means of transportation – does it get us to our destination faster than some other? We confront a colleague – did it change his or her behavior or make us feel better? In these and a thousand other instances, we tend to go with whatever we think works best.

What does it mean, though, to say that some approach “works” or that it works better than another does? Here, matters get a little complicated. Consider the toothpaste example. We may think it is a good idea to buy a special type of toothpaste, call it A. Should we buy it? Will we obtain a better result using it rather than our current toothpaste? Perhaps a researcher undertakes a study and shows that a group of people who used toothpaste A had fewer new cavities than they acquired on, say, an annual basis, prior to using it. That improvement seems self-evident to be a good thing, but can we trust that it resulted from using the toothpaste? No. Perhaps the improvement would have occurred regardless of the change in toothpaste. So, all that we can say is that an important outcome (fewer cavities) improved after using the new toothpaste.