This is a follow up to an earlier post: https://highresolve.wordpress.com/2014/05/23/are-racial-minorities-expelled-too-often-in-utah-schools/.
Any race-based study is complicated by the possible existence of other factors that might better explain any observed disparity. For instance, one might argue that, in reality, the disparity between disciplinary actions among the races might be better explained by socioeconomic status. To evaluate this possibility, we used the percent of students enrolled in the free or reduced school lunch programs (FRPL). (NOTE: The use of the FRPL program as a poverty indicator is a well-documented and common practice. See, e.g., http://nces.ed.gov/programs/coe/indicator_clb.asp; http://www.nyskwic.org/get_data/indicator_narrative_details.cfm?numIndicatorID=31. But http://febp.newamerica.net/background-analysis/federal-school-nutrition-programs (noting that it is generally a reliable indicator at the elementary school level, but that it may underestimate poverty at the high school level).)
We compared the Chi Squared value from our statistical analysis of disciplinary actions as an indicator of the relative racial disproportionality in disciplinary actions. We compared this value for each district with the percentage of students enrolled in the FRPL program for that district to create a scatter plot. Using linear regression analysis, we determined that the coefficient of determination, an indicator of the strength of the relationship between two variables, was 0.0218. A coefficient of .3 to .7 is the generally accepted range where a relationship is said to be recognizable. Here, the relationship between the variables is virtually nonexistent. In sum, we conclude that socioeconomic status can be ruled out because students of different races are subject to similarly elevated rates of discipline in both economically affluent districts and economically impoverished districts.