There is danger in how Murray makes statistical comparisons: for a valid relationship to emerge -- whether causal or correlation -- statisticians need to ensure that the observed sets of data are valid. Here's the experiment that is needed to disprove Murray's claims: collect two sets of data with similar attributes except for IQ which is treated as a random variable. E.g. one data set is black Americans, income bracket $50k-$100k, BA degrees, home ownership and mom & dad in home. The second data set is a copy of the first *except* white Americans (not Asian, not Hispanic). Run regressions to see whether the differences in IQ are random or attributed to ethnicity. Without such an experiment, how can anyone take Murray seriously? PS this experiment should be conducted by neutral, non-political economists, for obvious reasons.
There is danger in how Murray makes statistical comparisons: for a valid relationship to emerge -- whether causal or correlation -- statisticians need to ensure that the observed sets of data are valid. Here's the experiment that is needed to disprove Murray's claims: collect two sets of data with similar attributes except for IQ which is treated as a random variable. E.g. one data set is black Americans, income bracket $50k-$100k, BA degrees, home ownership and mom & dad in home. The second data set is a copy of the first *except* white Americans (not Asian, not Hispanic). Run regressions to see whether the differences in IQ are random or attributed to ethnicity. Without such an experiment, how can anyone take Murray seriously? PS this experiment should be conducted by neutral, non-political economists, for obvious reasons.