Having said that, here is what the authors found. They gathered several types of data and divided them into two discrete studies looking for specific correlations. They had 1,017 subjects across the two studies do various tests. One provided 10 topics, four of which were completely fake (such as subjunctive scaling, declarative fraction, genetic autonomy, neural acceptance), and then asked to rate their knowledge of each topic and tasked to create a convincing sounding explanation for each. This tested the subjects’ willingness to bullshit and ability to bullshit. The generated explanations were then given to other subjects to rate for accuracy and satisfactoriness. Subjects were then further given two measures of cognitive ability, the Wordsum task, which measure language ability, and the Raven’s progressive matrices, which measure abstract reasoning. Finally they were also tested for their receptiveness toward pseudoprofound bullshit, using methods similar to the study linked above. What they found in the first study was that, in line with their hypothesis, those who scored higher on bullshitting ability also scored higher on the two cognitive tests. So there was a modest correlation between intelligence as measured in this study and ability to bullshit. However, there was no correlation between ability to bullshit and willingness to bullshit (despite greater facility). And there was a negative correlation between intelligence and willingness to bullshit and receptivity to pseudoprofound bullshit.