| 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 Six:
Teacher Data
      The following data analyze several variables directly related and indirectly related to the teachers in Ohio’s districts relative to the variables of Presage Factor, percent passing, and actual performance. It should never go unnoticed that classroom teachers bear the brunt of the accountability effects of using OPT as an assessment mechanism for teacher effectiveness via the district ratings of OPT. Likewise, stakeholders, the media. and educational administrators who accept OPT and OSRC at face value make classroom teachers the target of their focus when angry or frustrated about low district scores.
      In my many and frequent discussions with classroom teachers, including those from districts where passing levels are above average, speak volumes to the problems OPT and OSRC have created for classroom teachers. Almost without exception, they articulate how OPT-driven management has taken from them the last vestiges of reflective, professional decision making about what is best for the children in their classrooms.
• Teacher Salary
      This graph shows the correlation between teacher salaries and the presage scores for the districts. In terms of the correlation coefficient (r=0.35), there is a moderately high degree of association between advantagement-disadvantagement and teacher salaries. Simply put, we clearly see that teacher salaries increase as a function of the wealth of the districts. In terms of school variables, as opposed to non-school variables, this finding is significant in terms of understanding additional apparent inequities across Ohio’s school districts. The finding here also underscores the problem of the spectrum of advantagement-disadvantagement when we consider the strong tendency for the more disadvantaged districts to have the most underpaid teaching staffs.
      The analysis of percent passing and teacher salary yields a moderately high correlation. This finding supports the notion from the graph of presage scores and teacher salary that advantaged districts tend to pay their teachers more than less advantaged districts. Likewise, the finding here supports the notion of OPT advantagement-disadvantagement bias because of the association of higher salary with higher percent passing. However, because percent passing is a function of OPT bias as established in the primary findings of this study, the claim that performance is a function of salary may be misleading.
      In examining actual performance we see a slight correlation between teacher salary and performance. This finding suggests that to some degree, teacher salary is a positive school variable. We must remember that actual performance is derived from controlling only for the effects of Presage Factor. The finding is not absolute because we cannot declare actual performance to be a robust measure of real academic performance beyond its presage control. However, relative to the other variables run against actual performance, teacher salary has the highest correlation with the exception of extra academic performance which will be presented later in this section.
• Degree Status
      The following sets of graphs examine district teacher degree status, the percent having no degree, bachelor’s degree, and master’s degree or higher.
Non-Degree:
      The analysis of the association of presage scores with the percent of teachers without a degree shows us that the correlation is not significant. However, the association of non-degree teachers tends to increase with greater disadvantagement.
      The graph of the analysis of percent passing with non-degree teachers yields what we would expect given the finding of the Presage Factor. Since percent passing is so closely defined by presage effects, this graph supports the findings in the first graph of this set.
      Controlling for presage effects, this analysis of percent non-degree teachers across actual performance yields a non-significant correlation because r=0.02 is extremely low. What minuscule correlation there is relates positively to increased actual performance. However, no claim to any statistical significance can be made.
Bachelor’s Degrees:
      The following graphs and analyses are best understood when examined in conjunction with the master’s degree graphs and findings because percent of teachers with master’s degrees and percent of teachers with bachelor’s degrees are fundamentally complimentary numbers, excluding the small percent of non-degree teachers. In other words, for any given district, the number of non-degree, bachelor’s degrees, and master’s degrees held by the teachers equals 100%.
The graph of presage score and teachers with bachelor degrees shows a slight inverse correlation. This finding tells us that the percent of bachelor degrees decreases somewhat as advantagement increases. This finding in and of itself may be seen as somewhat puzzling until we examine the finding regarding the percent of teachers with master’s degrees or higher. (See the third set of graphs in this section.) Taken with the findings of the analysis of master’s degrees and presage scores, the conclusion is that as advantagement increases so does the percent of teachers with master’s degrees or higher.
      The findings of the analysis of percent passing and teachers with bachelor’s degrees clearly show the artifacts of OPT bias along the advantagement-disadvantagement continuum. The correlation (r-0.23) is moderately significant and does show the tendency of wealthier districts to have greater numbers of teachers with degrees beyond the bachelor’s level when we consider this finding along with the finding regarding master’s degrees as discussed above.
      Again, taking the finding of actual performance and teachers with bachelor’s degrees along with its complement of teachers with master’s degrees or beyond seen below, we find that actual performance does correlate inversely, though only slightly. (Refer to the discussion following the presentation of the graph showing actual performance and teachers with master’s degrees or higher for more interpretation of this analysis.)
Master’s Degrees:
      The following graphs and analyses are best understood when examined in conjunction with the bachelor’s degree graphs and findings because percent of teachers with master’s degrees and percent of teachers with bachelor’s degrees are fundamentally complimentary numbers excluding the small percent of non-degree teachers. In other words, for any given district, the number of non-degree, bachelor’s degrees, and master’s degrees held by the teachers equals 100%.
      As we would expect, the percent of teachers with master’s degrees and beyond increases as a function of increase in advantagement. This finding is understandable in light of the extra expenditure required for paying salaries of teacher’s with graduate degrees.
      The finding of a moderate correlation between percent passing and percent of teachers with master’s degrees is not unexpected given the very high correlation (r=0.80) between percent passing and presage scores. In other words, districts with greater advantagement are more likely to have more teachers with advanced college degrees than those with less advantagement.
      The comparison of actual performance to the percent of teachers with advanced degrees shows a slight positive correlation. The analysis here tells us that within the previously discussed limits of actual performance in controlling for the presage effects, teacher’s having advanced degrees does contribute somewhat to actual performance.
• Teacher Experience
      The next section deals with the analysis of teacher experience across the variables of presage score, percent passing, and actual performance.
      The correlation between presage score and years of teacher experience is non-significant, though it shows a very slight inverse correlation that says there is a very slight tendency for more advantaged districts to have teachers with slightly less experience than less advantaged. However, the association is too low to support any claim other than there is no significant difference in the average years experience across Ohio’s school districts in terms of advantagement-disadvantagement.
      The finding from the analysis of percent passing and teacher experience shows an almost perfectly random correlation. In other words, there is no difference whatsoever in terms of teaching experience and percent passing OPT.
>
      The analysis of actual performance and teaching experience shows a slight positive correlation. Though the correlation is slight, it nonetheless appears to be a possible contributor to academic performance when we control for the effects of advantagement-disadvantagement.
Related Variables:
      Two variables related to teachers and teaching have been included in this section for possible illumination of the study’s findings. They are class size and extra-academic opportunities.
• Class Size
      The EMIS provides two similar sources of information regarding teacher/student ratios, class size and teachers per 1000 students. Both variables yield almost exactly the same findings. Since class size is a more familiar concept than teachers per 1000 students, I chose to use it.
      Class size is found to be inversely correlated to presage scores and is only slightly significant. This finding simply tells us that class size tends to be slightly lower the more advantaged the district is in terms of presage scores.
      Though slightly lower in terms of its statistical correlation, percent passing and class size reiterate the relations between presage effects and percent passing.
      Analyzing class size and actual performance yields a non-significant correlation that approaches randomness in the relations between the two variables.
• Extra Academic Opportunities
      The state defines extra academic opportunities as extracurricular activities that are academic in nature, such as debate team, French club, math club and similar activities open to student involvement outside the regular academic classes. Recreational and sports activities are not considered as extra academic opportunities.
      Analysis of the variables of extra academic opportunity and presage scores yields a moderate correlation. This means that extra academic opportunities increase as the advantagement as measured by the presage score increases.
      The correlation found with extra academic opportunities and percent passing is moderately high and tells us that the districts with greater numbers of such opportunities tend to perform better on OPT. However, it is important to remember the bias of OPT toward more advantaged districts. Because of this, conclusions regarding the actual effects are somewhat unclear, but the suggestion that extra academic opportunities contribute to improving district test performance is evident.
      The examination of extra academic opportunity and actual district performance shows a moderate correlation. This finding lends strength to the power of such opportunities in affecting actual test performance within the parameters of the Presage Factor. As well, this finding suggests that the idea discussed in the previous graph that extra academic performance positively affects percent passing may have greater credibility.
• Teacher Data Comments:
      Examination of the analyses and findings regarding the variables within this section on teachers as they may interrelate, indicates that most of the results are verifications of what we might expect given the power of the non-school forces and factors associated with district levels of advantagement-disadvantagement as described by the Presage Factor. However, if we can accept that actual performance is indeed a usable measure of what may be happening academically in Ohio’s schools, several findings in this section suggest themselves as variables contributing to that performance.
      Teacher salary, having a master’s degree or higher, years of teaching experience, and extra academic opportunities stand out as variables contributing to some degree to actual district performance. To examine further the efficacy of these variables, the four were converted to z-scores, added, and averaged to create a single measure. For lack of a better term, the combination into a single variable is called the "Teacher-Curriculum" variable (TC) to represent the domains of schooling from which the variables arise. The following three graphs examine the TC variable to better understand the potential of the four elements operating together.
      The correlation with presage scores is moderate and suggests that the variable is associated with the advantagement defined by the presage scores.
>
      The teaching-curriculum variable attains an moderately high correlation when associated with percent passing. The degree of association is higher than with the presage scores as seen in the preceding graph, thus suggesting it is more closely associated with percent passing than with presage scores themselves.
      Whereas three of the four variables that comprise the TC variable have less than r=0.15 correlation and the fourth variable of extra academic opportunity has a correlation coefficient of r=0.24, the combination of all four exceeds the average of the four coefficients. Again, it is important to remember that actual performance is a measure of OPT performance controlling for presage effects that represent the overwhelming bias of OPT. Another way to view this is to think of actual performance scores as possibly valid representations of academic performance. If this assumption is true, then the formulation of the TC variable begins to reach into the myriad of complex possibilities that shape authentic academic performance.
      The point here is not to posit a new way to assess district performance, but to demonstrate how complex the processes of teaching, schooling, and learning are in the real world of education. More specifically, the findings of this study to this point do not only tell us that OPT is a highly invalid assessment mechanism, but the findings also expose the tremendous complexity and difficulty of authentically and validly assessing academic performance on any level.
      Even if the calculus used to formulate actual performance results in a valid assessment of district performance to a greater degree than does OPT, the problem still exists that it is based upon a high stakes test. The pressure to pass the test, the time spent practicing to take the test, and the denial of the credential of a high school diploma for those innocent children who often narrowly fail to make the cutoff score is all born by the children and parents of Ohio’s public school population. Thus, there is no suggestion whatsoever that the derived actual scores should in any way be used to hold anyone accountable because of the damage such testing does to the children and to the curriculum they should have the opportunity to experience.
| 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 |