Social science has undergone a “credibility revolution,” prioritizing research designs that isolate cause and effect over mere correlation. The goal is to identify the impact of any specific “treatment” by approximating a counterfactual—what would have happened without it. This is often achieved through randomized controlled trials, as seen in a study by Esther Duflo, Michael Kremer, and Jonathan Robinson that offered Kenyan farmers “small, time-limited discounts on the cost of acquiring fertilizer (free delivery) just after harvest” (Duflo, Esther, et al. “Nudging Farmers to Use Fertilizer: Theory and Experimental Evidence from Kenya.” National Bureau of Economic Research, July 2009). By subsidizing purchases of fertilizer just after harvest, when farmers wanted to improve soil but didn’t have money left over to do so, the carefully timed intervention revealed a causal link between the nudge and increased fertilizer use: Farmers who had access to the subsidies were up to 20% more likely to fertilize their fields.

Naked Statistics equips readers with the tools for this kind of thinking. Wheelan presents regression analysis as a method for statistical control, but his core message aligns with the credibility revolution. He stresses that “To understand the true impact of a treatment, we need to know the ‘counterfactual,’ which is what would have happened in the absence of that treatment or intervention” (240). While regression can adjust for confounding variables, its power is limited by issues like omitted variables and reverse causality. Therefore, Wheelan champions experimental and quasi-experimental designs—like the Project STAR class-size study, the natural experiment of policing on high-alert days, and discontinuity designs in juvenile sentencing—because they generate more reliable counterfactuals.