| Home |
| 1-Overview || 2-FAQ || 3-Primary Findings || 4-Actual Performance | | 5-Funding Variables |
| 6-Teacher Data || 7-Race || 8-OSRC || 9-Closing Statement || Appendix-Top Performing Districts |

Forces and Factors Affecting Ohio Proficiency Test Performance:
A Study of 593 Ohio School Districts

Randy L. Hoover, Ph.D.

Section Seven
Percent African-American and Percent White as Variables
Across Presage Score, Percent Passing, and Actual Performance
      The following are three sets of graphed data addressing how OPT may be seen to play out across African-American and White school district populations. The first set of graphs below compares the association between the Presage variable and the percent African-American and the percent White of the district student population. The second set of graphs represents the percent passing per district as a function of percent African-American and percent White student populations. The third set gives a comparison of performance controlling for the Presage variable. It is vital that these three sets of findings be viewed together for optimal understanding of the general comparative effects of OPT on these two groups.

      The first graph in the set tells us that there is a moderate negative correlation between advantagement and the percent African-American student population. The second one shows that there is a moderate positive correlation between advantagement and the percent White student population of the district. The graphs are essentially inverses of each other as would be expected because as percent White goes up, the percent of African-American must go down and vice versa.
      Most simply stated, these graphs tell us that the greater the White population of the school district, the greater the level of advantagement; the greater the African-American district population, the greater the level of disadvantagement. Taken together, the findings support the argument that the effects of social-economic advantagement-disadvantagement are seen to a moderate degree in the racial composition of Ohio's school districts. The next set of graphs represents the findings of how the two populations are associated with OPT performance.

      This set of graphs shows us that there is a moderately significant differential between African-American and White performance on the OPT. The first graph in the set tells us that the greater the percent African-Americans in the district, the more likely fewer students will be achieving passing OPT scores. The second one shows the opposite for White students. Again, the graphs are basically mirror images of each other, as noted previously.
      Two points for objective interpretation are very important here. The first point is that the findings definitely reveal OPT bias against students in predominantly African-American school districts. It is also logically true that the findings may be interpreted as definitely revealing OPT bias in favor of predominantly White school districts. However, the second important point is that while the effects are real in terms of bias, it cannot and must not be concluded from this data array that the OPT bias is caused directly by racial differences between the two groups.
      While the findings do show the racial bias to be real, attributing the bias to a specific source requires more critical examination of the data. This is so because the study’s primary and most powerful finding is that social-economic advantagement-disadvantagement is the most significant predictor of performance. In other words, the research question arises of whether the demonstrated OPT bias shown here is a function of social-economics or race, or both.
      Indeed, it is somewhat interesting that the correlation coefficients for the first two sets of graphs are quite similar (r=-0.34, r=0.030 and r=-0.35, r=0.31). Considering the fact that we know from the primary findings of this study that the Presage factor is unusually powerful as a variable (r=0.80) of advantagement-disadvantagement for predicting OPT performance, to determine first-order racial/cultural OPT effects, we need to examine actual OPT performance controlling for the Presage factor. In other words, the racial differential shown in this second set of graphs must be examined further before suggesting that it is caused by racial/cultural differences and not by the social-economic factors of the Presage variable.
      The graphs below show actual performance (performance controlling for the Presage variable) for White and for African-American populations and yield statistically non-significant effects in and of themselves. These non-significant effects are, however, very significant in understanding and knowing that when the factors of advantagement-disadvantagement as defined by the Presage Factor are removed, we find that race is not the primary factor affecting academic achievement in terms of district level OPT performance. However, given the correlations (r=-0.15 for African-Americans and r=0.11 for Whites), we do see effects that could very well represent racial/cultural bias inherent in the tests.

      Though the correlations are low, the question does arise as to the source of why there would be any difference between African-Americans and Whites across OPT performance when controlling for the Presage Factor. Though nothing definitive about the primary source of the differential is immediately apparent, I would suggest two possibilities be given consideration.
      The first possibility is that there are other social-economic effects showing up that are not within the scope of the Presage Factor that are experienced differentially by the two groups such as is seen in the group correlations in the first set of graphs. The second possibility and certainly the more serious of the two is that OPT contains significant racial/cultural bias.
      My best professional judgment tells me that it is quite likely that the findings shown in the last set of graphs are related to racial/cultural OPT bias. I base this speculation on my experience and my intuition combined with the language/reading dependency of the test. Minimally, these data call for a complete and thorough examination of OPT for racial/cultural bias if the State of Ohio wishes to make any claims of test validity.

Summary Reflections:

• Together, all three sets of comparisons re-confirm that the social-economic-environmental factors that shape the conditions of advantagement-disadvantagement are the clear bias of OPT regardless of race. In other words, disadvantaged children are likely to perform more poorly on the test than advantaged children regardless of whether they are African-American or White.
• The first two sets of graphs do tell us that African-American children are more likely to suffer from the conditions of disadvantagement than are White children and that because of this powerful effect, far more African-American children are victims of OPT bias than are White children.
• The last set of graphs suggests that there is most likely some racial/cultural bias in the test. Though the effect of this bias on district level performance is significantly less than the effect of the Presage factor, it does make a difference especially in districts with high African-American student populations because passing rates load so powerfully on the Ohio School Report Card ratings.
• If indeed, the third set of graphed data is showing the artifacts of inherent racial/cultural bias, then the effects are particularly significant for individual African-American students in Ohio's schools. My own professional judgment tells me that this scenario is probable given the legacy of racial/cultural bias in standardized testing. At the very least, these data indicate a moral responsibility and a legal obligation on the part of the State of Ohio to suspend testing until further study of the racial/cultural bias is openly and honestly conducted.
• It is vital to recognize that the data represented in all three sets of graphs removes from discourse and discussion the question of African-American students ability to learn as well as other racial groups. There is simply no evidence whatsoever to support any arguments regarding inferior academic performance. Therefore, for anyone to make the claim that African-American children do not have the same native ability as Whites in terms of academic achievement is as absurd as it is ignorant and racist.
• Given the previous point, it would be remiss to fail to point out that OPT district performance as reported by the state (See, "Percent Passing and Percent African-American Students" in the second set of graphs in this section) makes it appear that African-American students are academically inferior to White students. The findings of this study do not support this implicit claim made by the State of Ohio through the OPT and OSRC. Indeed, the data indicate the claim is totally false and dangerously misleading in its racial significance.
• Again, the findings reported in this section of the study indicate a moral responsibility and a legal obligation on the part of the State of Ohio to suspend proficiency testing until the possible racial/cultural bias is thoroughly examined and the misleading racial overtones found in the results are corrected.
| Home |
| 1-Overview || 2-FAQ || 3-Primary Findings || 4-Actual Performance | | 5-Funding Variables |
| 6-Teacher Data || 7-Race || 8-OSRC || 9-Closing Statement || Appendix-Top Performing Districts |