Second, I agree that quantitative researchers look for non-linearity, but I think G&M are pointing out that even when they do, they tend to do it in smooth ways – age squared, log GDP, and so on. Given the focus on identifying average effects generated by within-sample variation (e.g. moving from one standard deviation below to above the mean on some independent variable), this makes sense. But it mostly excludes a search for mass points: the sharp non-linearities in the data corresponding to conceptual categories which are the bread and butter of qualitative research.