Whatever the reason, the regression that generates Lott and Mustard's stunning crime-reduction figures yields other results that might raise one's eyebrows. For example, Lott and Mustard examine two variables that seem to distinguish urban from rural counties-total population and population density. Both variables, however, are negatively correlated. with murder, rape, and burglary rates. 9 Perhaps, other things being equal, one is safer from murder, rape, and burglary in Chicago than in Valparaiso. More probably, however, in the words of O.J. Simpson's forensic expert Dr. Henry Lee, "Something's wrong. "
Similarly, the Lott and Mustard paper reports that, as one would expect,the proportion of a county's population that is black, male, and between the ages of ten and nineteen is strongly (and positively) correlated with the rate of property crime." When one turns to Lott and Mustard's tables, however, one learns more interesting things. Contrary to ordinary expectations, the proportion of a county's population that is black, male, and just a bit older is negatively correlated with the rate of property crime.' Moreover, the proportion of
the population that is black, male, and between ten and nineteen is less significantly correlated with the murder rate than is the proportion that is black, female, and over sixty-five. At the same time, the proportions of these two high-risk groups (teenage black males and elderly black females) are less powerful predictors of the homicide rate than the proportion of black women between forty and forty-nine. Jens Ludwig, noting similar anomalies in other crime categories, observes that either middle-age and retirement-age black women play a greater role in murder, assault, and auto theft than most of us have suspected or Lott and Mustard's models are misspecified.Once more, Dr. Henry Lee's immortal words might ring in one's ears: "Something's wrong”
ALBERT W. ALSCHULER
http://scholar.valpo.edu/cgi/viewcontent.cgi?article=1854&context=vulr