| 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 |
Randy L. Hoover, Ph.D.
Federal, State, and Local Funding Variables
This section addresses the role of federal, state and local funding percentages in terms of the Presage Factor, OPT performance, and actual OPT performance as defined in Section 4. Funding has historically been a source of contention across the arguments of stakeholders regarding school and district effectiveness. The major reason the arguments have continued is that there has not been a valid outcome measure of effectiveness against which to base the arguments. With the institution of OPT, critics of schooling in Ohio have used percent passing as the outcome measure to support their claims. In doing so, these critics have assumed that OPT is a valid measure of academic achievement and a valid measure of professional accountability.
It is now important to realize that the assumptions about funding using OPT as the bottom-line are invalid because of the bias found in this study. However, knowing the bias and being able to triangulate funding effects using presage scores, percent passing, and actual district performance does tell us what we claims cannot be made and does illuminate possible effects of funding beyond current public discussions.
Federal Revenue Effects
The first three graphs in the set below examine the percent federal funding a district receives and its association with district performance, Likewise, the subsequent sets of graphs examine state and local funding in the same manner.
This first graph above tells us clearly that federal funding is inversely correlated with advantagement-disadvantagement (r=-0.80). This correlation obviates the reality that disadvantaged districts are more eligible for federal funding than are more advantaged districts because of federal funding criteria.
Percent passing OPT and federal revenue show a significant negative correlation (r=-0.65), which tell us the greater the amount of federal funding, the lower the district OPT performance. Taken at face value outside the context of the studys findings, it would seem that federal funding has, at best, no effect on district performance and at worst a negative effect. However, this would be a superficial and terribly inaccurate interpretation given what we know about district OPT performance in relation to the elements of the Presage Factor.
Since we already know the bias of OPT against disadvantaged districts and because we know that disadvantaged districts receive far more federal revenue than advantaged districts, the results tell us nothing about the real effects of federal revenues. Thus, the primary finding here is that comparing federal revenue and district test performance tells us nothing about the effectiveness of the federal funding. Indeed, it is entirely possible, and I believe probable, that federal funding is tremendously beneficial to disadvantaged school districts.
Comparing actual performance as defined in Section 4 to federal revenue gives a correlation that is not statistically significant. However, the fact that there is no significant correlation considered along with the fact that the more advantaged districts do not receive any substantial federal funding suggests something potentially significant.
Remembering that actual performance controls for the bias of disadvantagement and that federal funding is a function of disadvantagement, a low correlation would be expected. However, given that approximately 50% of the districts falling within the range of disadvantagement are performing at or above what would be expected of them, it is entirely possible that this performance may be helped by the federal funding in those districts. We can make no claim about federal funding being ineffective from the findings and must reserve judgment for further study to determine potential positive effects in terms of actual test performance. In other words, without federal revenue, it is entirely possible that actual district performance would be far below what is seen for disadvantaged districts.
State Funding Effects
The next set of graphs does with state revenues what was done with the federal revenues discussed above. Likewise, the findings are somewhat similar to the findings for the effects of federal revenue.
Percent state revenue and presage scores have an inverse correlation (r=-0.51). Though less than the correlation found for federal revenue, the correlation is more than likely due to the same phenomenon. This is so because, despite the controversy regarding the inequity of Ohios funding formula, the state does fund in a compensatory manner relative to the advantagement-disadvantagement of the local districts. In general, under Ohios current funding formula economically disadvantaged districts do receive greater state subsidies than do economically advantaged districts.
Because state funding is not as compensatory as federal funding, more advantaged districts acquire more total funding when consideration is given to the additive factor of local funding. Therefore, the first graph suggests that, like federal funding, state aid may be viewed to a large degree as a surrogate measure of the presage variables. Therefore, on this basis, the inverse correlation is what we would expect to find.
Percent passing and its association with the percent of state funding, again, is what we would expect given the bias of OPT in favor of more advantaged school districts. Likewise, the correlation coefficient is lower than the federal one because Ohios school funding formula is less compensatory than the federal formulas for all programs combined. The finding must not be construed that state funding does not contribute to district effectiveness.
The correlation coefficient for actual performance (r=-0.10) and percent state revenue is not statistically significant. Just as with the interpretation given previously for the correlation of percent federal revenue, the effect of state funding on actual performance may be suggesting that it does have an effect that would more than likely be conspicuous by its absence but not provable by its presence because of OPTs lack of validity and the compensatory relationship of state funding to factors associated with the elements of the Presage Factor. In other words, without the existing levels of state funding, it is entirely possible that actual district performance would be far below what is seen for disadvantaged districts and possibly lower even for advantaged ones.
Local Funding Effects
The set of graphs representing local revenue contributions stands in striking comparison to the two previous sets of graphs in their positive correlations. However, the same principles of interpretation apply to these findings as do the previous findings for federal and state revenues. It must be remembered that in these three sets of graphs dealing with federal, state, and local funding, we are dealing with percent of total funding, so the data and graphs have the common denominator of being complements of each other in terms of the total federal, state, and local funds equaling 100%.
The correlation between presage scores and local revenue shows essentially what we would expect. Because district local revenue is a function of advantagement-disadvantagement with the percent of local funding increasing proportionally with the wealth of the district, this graph simply exposes the degree of that relationship. The data also does confirm the importance of understanding the power of local economics as related to school funding and advantagement-disadvantagement. Likewise, the finding also may point to possible questions of inequity in Ohios school funding program.
The correlation coefficient for percent local revenue and percent passing is, again, showing local revenue in its association with advantagement-disadvantagement given what know about the correlation of percent passing to presage effects and about local revenue being a function of advantagement-disadvantagement, as discussed previously. Taken only on its face, there appears to be the basis for the argument that increasing percent of local funding causes higher rates of passage. This claim of causality is not supportable given what we already know about the associations of higher levels of local funding being significantly correlated with higher levels of advantagement.
Though there is a slight positive correlation between local revenue and actual district performance, it is not statistically significant. However, this finding lends support to the idea that actual performance as defined by controlling for presage effects is stable across the variables of federal, state, and local funding contributions. This stability across variables reinforces the proposition that the Presage Factor taps into a robust measure of predictive validity for district performance and also supports further the idea that actual performance is likely a function of school variables as opposed to non-school variables.
Additional Local Funding Variables
The graph of residential valuation as a percentage of total valuation shows a moderate correlation with presage scores. However, a visual examination of the data plots shows a rather non-linear pattern that is also not curvilinear in shape. The upper right quadrant representing both higher presage scores and a higher percent of residential valuation does have a visual linear shape. The finding here suggests that from what might be called the "middle class" and upward to the more advantaged districts have greater yield from residential valuations than do most districts. The spread in the lower end of the presage scores, although not analyzed separate from the other data, appears to be rather random with the lower left quadrant being those districts with low property values.
The correlation and shape of the data plots for percent passing and residential valuation are consistent with the correlation and shape of the graphing of first graph of this set. This graph simply lends support to the notion of the positive correlation between district advantagement and test scores.
The actual performance and residential valuation graph findings support the idea that actual scores created by controlling for the presage effects stands up to testing against residential valuation. In other words, it helps support the hypothetical validity of actual performance.
The Related Variable of Per Pupil Expenditures
Though certainly not in the category of funding, per pupil expenditure in some ways can be seen to mirror funding and any discussion of revenue begs the question of spending. Therefore the variable of per pupil expenditure has been included in the funding section to provide additional insight with regard to how funds are expended in terms of Ohios pupils.
Per Pupil Expenditure
The correlation between presage score and per pupil expenditure is slight, though positive. Because the distribution of data is skewed away from linearity, interpretation is difficult at best. However, considering the slight correlation and the distributions in the upper-right and upper-left quadrants, the array does show a very uneven spread of per pupil spending that may indicate inequities in Ohio school district funding.
This graph clearly shows the general finding that there is little overall difference in the effects of per pupil spending except, again, in the upper right area of the data plots. This area clearly has higher per pupil spending, somewhat higher percent passing, and it is congruent with the outliers seen in the data plot in the first graph of this set. The finding lends some support to the idea that there are some very wealthy districts in terms of advantagement and that these districts are showing more clearly in this particular analysis.
Per pupil expenditure when viewed controlling for advantagement-disadvantagement supports the power of the Presage Factor as a relatively valid and stable indicator of OPT bias. This is so because the correlation coefficient (r=0.04) is extremely low thus indicating a non-significant correlation even lower than the correlations of the first two graphs in this set.