Suppose we have a dataset on the employment and labour market for the past ten years. You run a regression on the relationship between median wages (Y) and time (X) to investigate whether salaries have been increasing over the decade. You nd that time is positively associated with wages, i.e., the median wage has indeed been increasing. Your friend is curious to know whether this trend holds across all educational subgroups. There are four subgroups in your dataset: (i) secondary school dropouts, (ii) secondary school graduates with no tertiary education, (iii) those with tertiary education and (iv) those with Bachelor's and higher degrees. To your surprise, you nd that within all educational subgroups, the median wage is lower now than it was at the beginning of the decade. 0 How can both observations be true: overall salaries have risen, but the median wage within every educational subgroup has fallen