A June 8 San Francisco Examiner article discussed measures at the California state and San Francisco local levels aimed at addressing large racial differences in public school discipline rates by generally relaxing the stringent discipline policies in effect in recent decades. The thinking underlying the measures accords with the near universal perception, promoted by the U.S. Departments of Justice and Education and much commentary, that stringent discipline policies tend to cause large racial disparities in discipline rates. Exactly the opposite is the case. Stringent discipline policies tend to yield smaller racial differences in discipline rates than more lenient ones.
Inherent in the shapes of normal distributions of factors associated with experiencing an outcome is a pattern whereby the rarer an outcome the greater tends to be the relative difference in experiencing it and the smaller tends to be the relative difference in avoiding it. The pattern can be readily illustrated with data on employment tests on which some minority groups have lower average scores than whites. Lowering cutoffs has long been regarded as a means of reducing the racial impact of such tests because lowering cutoffs tends to reduce relative differences in pass rates. For example, if pass rates are 80 percent for whites and 63 percent for minorities, the minority pass rate is about 21 percent lower than the white pass rate. If the cutoff is lowered to the point where 95 percent of whites pass, assuming normal test score distributions, the minority pass rate would be about 87 percent. Thus, with the lower cutoff, the minority pass rate would be only 8.4 percent lower than the white rate.
But whereas lowering cutoffs tends to reduce relative differences in pass rates, it tends to increase relative differences in failure rates. In the situation just described, the minority failure rate was initially 1.85 times the white failure rate (37 percent/20 percent). With the lower cutoff, the minority failure rate would be 2.6 times the white failure rate (13 percent/5 percent).
This pattern is not peculiar to test score data or the numbers I chose to illustrate it. It will be found in almost any situation where groups differ in their susceptibility to some outcome.
School discipline standards operate just like test cutoffs. Relaxing standards, by increasing overall rates at which students avoid discipline, will tend to reduce relative differences between the rates of avoiding discipline (the equivalent of passing the test), but increase relative differences in discipline rates (the equivalent of failing the test).
An increase in the racial difference in discipline rates in fact occurred a few years ago when the Los Angeles Unified School District adopted the School Wide Positive Behavior Support program to promote measures to improve student conduct without resorting to exclusionary punishments like suspension. The year before the program was implemented black and white suspension rates were 25.1 and 12.0 percent, respectively; the year after, these rates were 22.6 and 3.2 percent. Thus, as both black and white suspension rates declined, the ratio of the black rate to the white rate increased from 2.1 to 7.1.
In the Los Angeles situation, something seemed to be at work beyond the statistical forces I described above. Those forces by themselves ought not to have caused such a large increase in the ratio of suspension rates. Moreover, contrary to the described statistical pattern, the racial difference in rates of avoiding discipline also increased.
One possibility is that there existed substantial racial differences in types of offense and that the offenses in which black students are more commonly involved were of a type where suspension remained the usual punishment. Such possibilities must be borne in mind even when policy makers fully understand the pertinent statistical forces. Currently, however, from the Departments of Justice and Education on down, almost no one understands those forces. As far as I can tell, few even understand that lowering test cutoffs tends to increase differences in failure rates.
One unfortunate consequence of the mistaken perception that stringent discipline policies cause larger racial differences in discipline rates than more lenient ones is that the perception seems invariably to influence appraisals of the utility of such policies. If stringent discipline policies are important to maintaining environments where students are able to learn, they are probably more important in inner city schools than anywhere else. Thus, the value of such policies warrants rigorous examination. But the studies I have reviewed seem to begin with the assumption that stringent policies cause large racial differences in discipline rates and proceed from there to some very dubious reasoning as to why the data show that the policies do more harm than good. Of course, the fact that researchers believe that stringent discipline policies cause larger racial differences in discipline than more lenient ones by itself calls into question their interpretations of complex data even if the interpretations are uninfluenced by that belief.
James Scanlan is a Washington, D.C., attorney specializing in the use of statistics in litigation.
The Recorder welcomes submissions to Viewpoint. Contact Vitaly Gashpar at email@example.com.