Evaluation of Desegregation Funding Goals

November 15^th, 1996

Gregg Cannon, Ph.D.

Executive Summary

The three primary goals of desegregation funding are: 1) to maintain or promote student diversity; 2) to improve achievement for all students, regardless of race, ethnicity or gender; and 3) to ensure equal access to schools, programs and activities, regardless of race, ethnicity or gender. This evaluation was an attempt at determining how well Desegregation (Deseg) schools are doing relative to meeting these goals. Since at present there is very little or no data available on participation in school activities this initial evaluation was limited to addressing the first two goals. The wording of the first two goals implies both, a "snap-shot" comparison between Deseg and Non-Deseg schools, and a longitudinal comparison of Deseg schools in terms of academic achievement and student diversity.

The following is a comparison among the District’s original 21 Desegregation (Deseg) Schools and 70 Non-Desegregation (Non-Deseg) Schools. Both Deseg and Non-Deseg schools included only elementary and middle schools. Since all Alternative Education Schools were Non-Deseg they were excluded from the analysis. Deseg and Non-Deseg schools were compared on their demographic characteristics, achievement scores, and responses to the School Quality Survey over the past two years. The comparison between Deseg and Non-Deseg schools on measures of achievement, and responses to the School Quality Survey was done using both residualized and non-residualized scores. A residualized measure has had the variance accounted for by some other related variable or variables removed.

In Effective Schools research the variance removed is usually variance accounted for by some variable or variables over which the school has no control. These variables are usually related to the demographic make-up of the student population. Removing the variance accounted for by the demographics of a school, in effect makes the schools more comparable.

Maintaining or Promoting Student Diversity

For both school years (1994-95 & 1995-96) Deseg schools appeared to be more ethnically diverse than Non-Deseg schools. There was a smaller percentage of White/Anglo students and a larger percentage of minority students in Deseg schools compared to Non-Deseg schools. The only exception to this was the smaller percentage of Native American students in Deseg compared to Non_Deseg schools. However, analysis showed the percentage of Native American students in Deseg schools increased from 94-95 to 95-96.

To evaluate how Deseg funding, and enrollment diversity is affecting diversity appreciation required analysis to the Diversity Appreciation questions from the School Quality Survey. The data in these tables presented compelling evidence that Deseg schools are typically rated higher than Non-Deseg schools in terms of Diversity Appreciation. The removal of variance due to cost per student alone had a detrimental effect on measures of diversity appreciation in Deseg schools, implying that increased cost per student in Deseg schools has a positive effect on diversity appreciation.

Improve Achievement for all Students Regardless of Race, Ethnicity, or Gender

Although Deseg schools scored a little lower on all four subscales of the ITBS, the difference was not statistically significant. Similar results were obtained wwhen ITBS scores were broken down by ethnic group. One notable pattern was that both Native American and Asian American students’ ITBS scores increased from 94-95 compared to 95-96 in Non-Deseg schools while they decreased in Deseg schools over the same period. Results of the same analysis broken down by gender indicated that in general female students out performed male students in both Deseg and Non-Deseg schools except in mathematics.

In most instances when comparisons of ITBS scores were based on residualized scores or demographic matching, Deseg schools scored higher overall than Non-Deseg schools.

When comparisons were based on raw Essential Skills scores Deseg schools scored slightly lower than Non-Deseg schools. On the other hand, when comparisons were based on residualized scores, matching, or both, Deseg schools scored higher than Non-Deseg schools.

Whether comparisons were based on residualized measures or demographic matching, Deseg schools were affected more positively than Non-Deseg schools. Since the removal of variance due to both student demographics and cost per student increased performance more in Deseg schools compared to Non-Deseg schools, it stands to reason that this variance has a more detrimental effect on performance in Deseg schools. More specifically, the lower SES, higher mobility, and higher percentage minority found in Deseg schools has a detrimental effect on achievement performance. For this reason, removal of the variance due to these same demographic factors has a positive impact on achievement performance for Deseg schools. By the same reasoning, since removal of the variance due to cost per student alone also increased performance in Deseg schools, it follows that increased cost per student has a detrimental effect on academic achievement. Since it would be hard to argue that increased cost per student (e.g., more teachers) has a detrimental affect on academic achievement, the most that could be said is that cost per student (i.e., the way Desegregation funds are currently being spent) has a negligible impact on academic achievement.

Evaluation of Desegregation Funding Goals

The three primary goals of desegregation funding are: 1) to maintain or promote student diversity; 2) to improve achievement for all students, regardless of race, ethnicity or gender; and 3) to ensure equal access to schools, programs and activities, regardless of race, ethnicity or gender. This evaluation was an attempt at determining how well Desegregation (Deseg) schools are doing relative to meeting these goals. Since at present there is very little or no data available on specific programs and activities this initial evaluation was limited to addressing the first two goals. The wording of the first two goals implies both, a "snap-shot" comparison between Deseg and Non-Deseg schools, and a longitudinal comparison of Deseg schools in terms of academic achievement and student diversity.

The comparison among Deseg and Non-Deseg schools on measures of achievement, and responses to the School Quality Survey was done using both residualized and non-residualized scores. A residualized measure has had the variance accounted for by some other related variable or variables removed. In Effective Schools Research it is often desirable to remove variance that impacts on achievement scores. Usually the variance removed is variance accounted for by some variable or variables over which the school has no control. These variables are usually related to the demographic make-up of the student population. Removing the variance accounted for by the demographics of a school, in effect makes the schools more comparable. Removing extraneous variance that impacts on measures of interest is a method of statistically matching schools.

Table 1 shows a statistical profile of the District’s 21 Desegregation schools. In addition to the four demographic variables used in the residualization process the table includes variables such as attendance rate, student/teacher ratio, and cost per student to name a few. Table 2 shows the results of an Analysis of Variance (ANOVA) comparison between Deseg and Non-Deseg schools on the same variables presented in Table 1. Only three variables were found to be significantly different at (p < .01). The average number of students per teacher was significantly greater in Non-Deseg schools compared to Deseg schools. The average number of students per teacher with advanced degrees was also significantly greater in Non-Deseg schools. Finally, the average cost per student was significantly greater for Deseg compared to Non-Deseg schools.

Table 1

Identification and Statistical Profile of the

District’s 21 Desegregation Schools

1 All residualized effectiveness ratings have had the variance due to SES, %LEP, %minority, and %mobility removed. Effectiveness Ratings have a mean of 0 and a standard deviation of 1.0. Negative values indicate schools scoring lower than expected given the population of students they serve. Positive ratings indicate schools scoring higher than expected given the population of students the school serves.

2 Control schools were matched on the Stress Index which is calculated as the weighted average of the schools SES, %LEP, %minority, and %mobility. The Stress Index like the Effectiveness Ratings have a mean of 0 and a standard deviation of 1.0. Schools with larger positive values are schools with less stress while schools with larger negative values denote schools with more stress.

3 Residualized ratings where the only variance removed was accounted for by cost per student. Like previous effectiveness ratings¹ the mean is 0 and the standard deviation is 1.0.

· Larger values indicate greater agreement to positively worded questions on the School Quality Survey

· Larger values indicate greater agreement to positively worded questions on the School Quality Survey

1 All residualized effectiveness ratings have had the variance due to SES, %LEP, %minority, and %mobility removed. Effectiveness Ratings have a mean of 0 and a standard deviation of 1.0. Negative values indicate schools scoring lower than expected given the population of students they serve. Positive ratings indicate schools scoring higher than expected given the population of students the school serves.

2 Control schools were matched on the Stress Index which is calculated as the weighted average of the schools SES, %LEP, %minority, and %mobility. The Stress Index like the Effectiveness Ratings have a mean of 0 and a standard deviation of 1.0. Schools with larger positive values are schools with less stress while schools with larger negative values denote schools with more stress.

2 Control schools were matched on the Stress Index which is calculated as the weighted average of the schools SES, %LEP, %minority, and %mobility. The Stress Index like the Effectiveness Ratings have a mean of 0 and a standard deviation of 1.0. Schools with larger positive values are schools with less stress while schools with larger negative values denote schools with more stress.

3 Residualized ratings where the only variance removed was accounted for by cost per student. Like previous effectiveness ratings¹ the mean is 0 and the standard deviation is 1.0.

Achieving the First Two Goals of Desegregation Funding

1 Maintaining or Promoting Student Diversity

Assessing how well Deseg schools are doing at meeting the first goal of desegregation funding requires looking at Tables 3 through 5. For both school years (Table 3 & 4) Deseg schools appear to be more ethnically diverse than Non-Deseg schools. There is a smaller percentage of White/Anglo students and a larger percentage of minority students in Deseg schools compared to Non-Deseg schools. The only exception to this is the smaller percentage of Native American students in Deseg compared to Non_Deseg schools. However, analysis of Table 5 shows the percentage of Native American students in Deseg schools increased from 94-95 to 95-96, although not at the same rate as Non-Deseg schools. Tables 3 through 5 support the position that Deseg schools are both maintaining and promoting student diversity.

To evaluate how Deseg funding, and enrollment diversity is affecting diversity appreciation requires analysis of Tables 25 through 29 that compares responses to the School Quality Survey. In Tables 25 through 27 it can be seen that Deseg schools are rated higher than Non-Deseg schools in terms of Diversity Appreciation. In Table 25 it can be seen that Deseg schools were rated slightly higher than Non-Deseg schools on positively worded questions pertaining to Diversity Appreciation. Table 26 shows an even greater disparity between Deseg and Non-Deseg schools following the removal of variance due to SES, %LEP, %minority, and %mobility. Table 27 shows Deseg schools being rated higher than a matched group of 21 Non-Deseg schools. Finally, Table 28 and Table 29 show the results of a comparison between Deseg and Non-Deseg schools using residualized measures where the only variance removed was due to cost per student alone. In Table 28 when the variance due to cost per student alone was removed ratings on Diversity Appreciation were found to be reversed (i.e., Non-Deseg schools scored higher than Deseg schools). If the removal of variance associated with some variable increases performance, it must be the case that retaining the variance associated with the same variable would have a detrimental effect on performance. It follows then that the higher Diversity Appreciation ratings for Deseg schools are more strongly associated with higher cost per student. This is not to say that increased cost per student caused higher ratings on Diversity Appreciation, but simply that parents, students, and teachers of Deseg schools tended to score their schools higher on Diversity Appreciation, and they tended to have greater cost per student (correlational relationship). Although not conclusive, these results suggest that Deseg funding is promoting diversity enrollment and diversity appreciation in Desegregation schools.

2 Improve Achievement for all Students Regardless of Race, Ethnicity, or Gender

Table 9 displays results of an ANOVA comparison between Deseg and Non-Deseg schools on ITBS (NCE) scores. Although Deseg schools scored slightly lower, the difference was not statistically significant. Tables 10 through 12 show similar results broken down by ethnic group. One notable pattern in Table 12 was that both Native American and Asian American students’ ITBS scores increased from 94-95 compared to 95-96 in Non-Deseg schools while they decreased in Deseg schools over the same period.

Tables 13 through 15 show results of the same analysis broken down by gender. The results show that typically female students out perform male students in both Deseg and Non-Deseg schools except in mathematics, where male students tend to score higher than female students. Tables 16 through 19 display results of an ANOVA comparison between Deseg and Non-Deseg schools using the same types of standardized and residualized scores used to compare Deseg and Non-Deseg schools on responses to the School Quality Survey. In most instances whether comparisons were based on residualized scores or demographic matching, Deseg schools scored higher than Non-Deseg schools. The only statistically significant difference between Deseg and Non-Deseg schools occurred when comparing Deseg to 21 matched Non-Deseg schools using residualized scores where the only variance removed was due to cost per student (Table 19). This method of comparison produced significant differences in Usage and Expression and Mathematics with Deseg schools scoring higher on both subscales.

Tables 20 through 24 compare Deseg and Non-Deseg schools on Essential Skills scores using the same methods of comparison used to compare ITBS scores and responses to the School Quality Survey. The comparison seen in Table 20 based on raw Essential Skills scores show Deseg schools scoring slightly lower than Non-Deseg schools. In Tables 21 through 24 whether using residualized scores, matching, or both, Deseg school scores higher than Non-Deseg schools.

Whether comparisons were based on residualized measures or demographic matching, Deseg schools were helped more than Non-Deseg schools. Since the removal of variance due to both student demographics and cost per student increased performance more in Deseg schools compared to Non-Deseg schools it stands to reason that this variance has a more detrimental effect on performance in Deseg schools. More specifically, the lower SES, higher mobility, and higher percentage minority found in Deseg schools has a detrimental effect on achievement performance. For this reason, removal of the variance due to these same demographic factors has a positive impact on achievement performance for Deseg schools. By the same reasoning, since removal of the variance due to cost per student alone also increased performance in Deseg schools, it follows that increased cost per student has a detrimental effect on academic achievement. Since it would be hard to argue that increased cost per student (e.g., more teachers) has a detrimental affect on academic achievement, the most that could be said is that cost per student (i.e., the way Desegregation funds are currently being spent) has a negligible effect on academic achievement.

In summary, results of this statistical comparison between Deseg and Non-Deseg schools suggest that desegregation funding, analyzed in terms of increased cost per student has little relationship to student achievement. On the other hand, desegregation funding as measured by increased cost per student does appear to be positively related to diversity enrollment and diversity appreciation.

DESEG963/GEC

Top of Page