Menu
EPLC Conference
About the Center
Calendar
Center News
The Policy Forum
Policy Projects
Leadership Programs
EPLC Reports and Publications
Education Policy Information Clearinghouse
Sign Up for EPLC News Services
Technical Assistance
Pennsylvania Education Policy Letter
Home

Poverty, Race, Resources, Results
in the
Pittsburgh Public Schools


THE ACORN STUDY


A report from the
NATIONAL CENTER FOR SCHOOLS AND COMMUNITIES
FORDHAM UNIVERSITY

Copyright April 30, 2003

Introduction

At the request of the Pennsylvania Association of Community Organizations for Reform Now (ACORN), the National Center for Schools and Communities conducted an analysis of the relationships among inputs, resources, demographics, and outcomes in the Pittsburgh Public Schools). This report summarizes the results of that investigation.
We find moderate evidence of positive relationships between teaching resources and academic outcomes. We find strong evidence relating attendance and outcomes. We also find evidence of income and race based patterns in the distribution of educational resources for children attending Pittsburgh's public schools.


Methodology

This analysis is based on a data set of nearly 200 variables compiled from a number of sources. The bulk of the data is drawn from the 2000-2001 Pennsylvania System of School Assessment - School and District Profiles (Pennsylvania Department of Education. The African American enrollment by school as of October 2, 2000 is based on the Pittsburgh Public Schools Membership Report for 2001-2002 (PPS Office of Technology). Per school teacher data not available in the state reports were provided by PPS to Pittsburgh ACORN and include average years of service, average sick days, number of teachers with Bachelor's degrees, number of teachers with Master's degrees, number of teachers with Doctorates, and total number of teachers. These data are for the 2002-2003 school year. Their use in this analysis represents a compromise of availability over the ideal. Our assumption is that the staffing pattern in schools generally does not shift precipitously; we are, however, willing to check our findings against 2000-2001 data when they become available to the community. We note that the school profile data currently available on the PPS website is for 1998-1999.

For academic outcomes, we have used variables related to the Pennsylvania System of School Assessment tests in math and reading which are given in the fifth, eighth, and eleventh grades. To allow for comparisons among the dominant organizational formats in the system (elementary, middle, high), we have excluded eighth grade data from schools other than middle schools.

The PPS teacher data refers to schools which are not listed in the 2000-2001 state data; this disparity leads to occasional instances of missing data, which are factored out of the analysis as necessary. In the comparison of best and most poorly performing schools, zero indicates such missing data. A small number of schools appeared to be special-purpose and, therefore, have been excluded from the analysis. Generally, our calculations are based on a maximum of 59 elementary schools, 20 middle schools, and 11 high schools.

All findings are at the 95 to 99 percent significance level, which means there is only a one to five percent chance that they could be random. Occasionally, we refer to correlations with lower significance levels but only by way of suggesting the overall direction of a group of correlations that do meet the significance standard.

We have attempted to maintain a narrative tone that is accessible to the non-technical reader. We usually refer to correlations as relationships. In a positive relationship, as x increases, y increases and vice versa. For example, in some cases, schools with higher average teacher experience tend to have higher percentages of students testing at the top level. A negative (or inverse) relationship suggests that as x increases, y decreases and vice versa. For example, as the percentage of low-income children in a school goes up, the average experience of their teachers tends to go down.

For the statistically inclined, we present tables of association in the appendices.


Academic Outcomes As A Function Of Educational Inputs

Attendance:

In this and other analyses, we have found that attendance is a gatekeeper variable that can be viewed as both an input and an outcome. Of the variables we examined for this report, attendance is among the clearest predictors of both average test scores and the percentage of students in the two top scoring categories for the fifth, eighth, and eleventh grade math and reading tests. Specifically, there is a positive relationship between overall elementary school attendance and:

  • the percentage of students scoring in level four (top) of the fifth grade math test
  • the percentage of students scoring in level four (top) of the fifth grade reading test
  • the percentage of students scoring in level three of the fifth grade reading test

The positive relationships at the middle school level are even stronger. Attendance relates to:

  • the percentage of students scoring in level four of the eighth grade math test
  • the percentage of students scoring in level four of the eighth grade reading test
  • the percentage of students scoring in level three of the eight grade math test
  • the percentage of students scoring in level three of the eighth grade reading test

In high school, the positive relationship of attendance to higher scoring groups is similar:

  • the percentage of students scoring in level four of the eleventh grade math test
  • the percentage of students scoring in level four of the eleventh grade reading test
  • the percentage of students scoring in level three of the eleventh grade math test
  • the percentage of students scoring in level three of the eleventh grade reading test

In other words, in all these instances, the higher the average school attendance, the higher the percentage of students scoring in the top two levels tends to be.

With the exception of eighth and eleventh grade math level two (.363 and .565, no significance), there is an inverse (negative) relationship between attendance and the percentage of students scoring in the bottom two categories for math and reading tests (significant for seven of ten coefficients). That is, in all significant instances, lower attendance is related to higher percentages of students in the lowest scoring levels.

Attendance is directly related to schools' average test (scale) scores for the standardized tests at all three levels:

  • fifth grade math
  • fifth grade reading
  • eighth grade math
  • eighth grade reading
  • eleventh grade math
  • eleventh grade reading

Clearly, schools with better attendance, on average, perform better on the state tests. The comparison of top and bottom ranked schools in the final section further illustrates this conclusion.

Teachers:

Various studies suggest that teachers' experience is related to the academic success of their students. We found a number of relationships between average teacher experience per school and student performance. Average years of service is positively related to:

  • average school scale scores on fifth grade math test
  • average school scale scores on fifth grade reading test
  • average school scale scores (scale) on the eleventh grade math test
  • average school scale scores on the eleventh grade reading test

In other words, higher scores on these tests are associated with schools that have more experienced teachers.

Generally, the relationship between average years of experience and the percentage of students in the two lower scoring levels is negative; that is, as the percentage of experienced teachers decreases, the percentage of students with weaker scores increases. There are negative relationships with lower score levels for the following tests and levels:

  • fifth grade math level one
  • fifth grade reading level one
  • eighth grade reading level two
  • eleventh grade math level one
  • eleventh grade reading level one

The one exception to this trend is the relationship of average teacher experience to eleventh grade math level two, which is positive.

Average years of teaching experience per school is also positively related to higher percentages of students scoring in the upper two levels of standardized tests in the following instances:

  • the percentage of students scoring in level three of the fifth grade math test
  • the percentage of students scoring in level four of the fifth grade reading test
  • the percentage of students scoring in level four of the eighth grade math test
  • the percentage of students scoring in level four of the eighth grade reading test
  • the percentage of students scoring in level three of the eleventh grade math test
  • the percentage of students scoring in level four of the eleventh grade reading test
  • the percentage of students scoring in level three of the eleventh grade reading test

Master's degrees:

Some research, particularly that cited by No Child Left Behind program staff, emphasizes the level and quality of teacher education. The data used for this study suggest a modest positive relationship between the percentage of teachers with master's degrees and average scale scores on both the elementary math and reading tests:

  • fifth grade scale score math
  • fifth grade scale score reading

There is also a negative relationship between the percentage of students testing at the lowest levels in math and reading and master's degrees; that is, the fewer the teachers with master's degrees, the more students testing at the bottom level:

  • fifth grade math level one
  • fifth grade reading level one

Educational program components:

The State of Pennsylvania collects data on a standardized set of programmatic resources. These resources are organized by level and categorized as academic or supportive. Academic options range from "art instruction with a certified art instructor" to tutorial programs to career exploration to honors courses. Supportive activities range from before school clubs to intramural sports to band to parent involvement. For elementary schools, the state tracks 15 academic programs and 12 supportive programs. For middle schools there are 26 academic and 11 supportive categories. High schools have up to 28 and 13, respectively. (See the appendix for a full list of activities.)

We constructed six rudimentary variables by totaling the number of state standardized academic and support activities at the elementary, middle, and high school level. The number of academic program activities an elementary school has a positive relationship to:

  • percentage of students at level four on the fifth grade math test
  • percentage of students at level three on fifth grade math test
  • fifth grade scale scores for math
  • fifth grade scale scores for reading

Of the various academic options, the presence of the elementary enrichment programming relates positively with the percentage of master's degrees. In other words, a school with a higher percentage of teachers with master's degrees is more likely to have this activity.

At the middle school level, the number of support activities relates positively to attendance, and, as we know, attendance relates positively to test outcomes.


Resources As A Function Of Demographics

Across all levels, there is a moderate negative relationship between the percentage of master's degrees at a school and the low-income enrollment: more poor students, proportionately fewer master's degrees.

Across all levels, there is a smaller but still definite negative relationship between master's degrees and the percentage of African American enrollment.

At different levels, there are negative relationships between the percentage of low-income enrollment in a school and:

  • the percentage of master's degrees in elementary schools
  • the average years of teaching experience in elementary schools
  • the percentage of master's degrees in middle schools
  • the average years of teaching experience in the middle schools
  • the average years of teaching experience in the high schools

In short, at every level, as the percentage of low-income children goes up in a school, the qualifications of its faculty tend to go down.

Moreover, there are also negative relationships between the percentage of African American children in a school and:

  • the percentage of master's degrees in elementary schools
  • the average years of teaching experience in the elementary schools

At the elementary level, there is a negative relationship between the percentage of low-income students and the available academic options; in other words, the poorer the school, the fewer state-defined academic programs available.

Of the various academic options, the elementary enrichment program relates negatively with the percentage of low-income enrollment.

The likelihood that an elementary schools has a certified foreign language program decreases as low-income enrollment goes up.


Schools Compared By Reading Rank

The following table compares the top and bottom ten percent of Pittsburgh elementary schools in terms of reading scores on the fifth grade test. In addition, we juxtapose the top and bottom four middle schools and three high schools ranked for their reading scores. This exercise provides a visual sense of the tendencies described in the previous statistics. Although any given school can vary from the trends, the spreadsheet provides "eyeball analysis" verification of those trends and adds the reality of putting the "face" of real Pittsburgh schools on the abstractions of our earlier discussion.

Five of the six top elementary schools also rank in the top twenty in terms of master's prepared faculty. Four out of five of the bottom six rank in the last twenty for master's degrees. (A zero indicates missing data.) Three of the high-end elementary schools have faculties in the top ten for elementary schools; on the other side of the comparison, three of five are among the ten least experienced faculties. The low-income enrollments of the top six are among the lowest in the city; at the other end are schools from among the twenty poorest. While somewhat subtler than these comparisons, the difference in the attendance rates for the two groups is still noticeable.

Three of the four top ranking middle schools have faculties in the top six for both experience and master's degrees. The differences in attendance rates between high and low ranked middle schools are more pronounced than those for elementary schools.

The three top reading high schools have the three highest concentrations of master's degrees and two of the three most experienced faculties. The polarization of higher and lower low-income enrollment is again obvious. Again, attendance patterns are distinct at the two ends of the scale.


School ZIP School Name
Reading rank
Experience rank
Master's rank
%AfrAm
AA rank
% Low Inc
L Inc rank
Attendance
Math rank
15201 Sunnyside Elem School
1
6
11
60%
27
54%
46
95%
33
15210 Roosevelt Elem School
2
21
6
4%
58
57%
43
92%
9
15211 Whittier Elem School
3
4
38
33%
46
61%
38
94%
19
15216 Beechwood Elem School
4
1
3
18%
50
47%
50
93%
10
15210 Bon Air Elem School
5
48
17
9%
56
29%
59
93%
1
15208 Linden Elem School
6
27
18
52%
34
40%
55
96%
4
15208 Belmar Elem School
54
45
42
99%
7
82%
17
93%
59
15219 McKelvy Elem School
55
0
0
100%
1
92%
1
91%
58
15212 Mann Elem School
56
26
55
39%
41
79%
20
94%
54
15221 Crescent Elem School
57
50
40
100%
5
85%
13
90%
53
15201 McCleary Elem School
58
51
5
38%
42
87%
10
91%
48
15214 Northview Heights Elem School
59
52
50
95%
11
92%
3
91%
52
15208 Sterrett Classical Academy
1
1
1
51%
14
36%
20
93%
1
15226 West Liberty Classical Academy
2
4
2
54%
9
62%
13
93%
3
15206 Rogers Ctr Creat & Per Arts
3
8
11
52%
13
38%
19
94%
4
15213 Frick Intl Studies Academy
4
6
6
53%
10
47%
16
92%
2
15212 Columbus Middle School
18
18
18
90%
4
85%
5
80%
17
15201 Arsenal Middle School
17
16
16
83%
5
89%
2
87%
16
15219 Milliones Middle School
19
7
8
99%
1
89%
1
77%
19
15219 Letsche Ed Ctr
20
3
10
79%
6
40%
18
68%
20
15208 Pittsburgh HS Creat & Per Arts
1
1
1
38%
8
32%
10
88%
3
15213 Schenley High School
2
3
3
64%
4
35%
9
83%
2
15217 Allderdice High School
3
5
2
28%
10
22%
11
88%
1
15212 Oliver High School
9
10
10
69%
3
56%
3
71%
8
15208 Westinghouse High School
10
8
4
100%
1
58%
2
71%
10
15203 South Vo-Tech High School
11
11
5
59%
5
63%
1
74%
11

 

Policy Questions

The description painted by the quantitative analysis of outcome, demographic, and resource data for the Pittsburgh Public Schools suggests a number of qualitative questions for exploration by parents, community groups, educators, and policy makers.

Do the patterns of resource distribution suggested by this analysis represent explicit policy on the part of the educational and political leadership of Pittsburgh? If not, how willing are that leadership and the community to alter those patterns? What policy changes will be necessary to change the current distribution essential educational inputs?

Specifically, how can school administrators persuade more experienced, better educated teachers to work with students most in need of their capabilities?

Given the demonstrable importance of attendance to academic success, is the educational program of the Pittsburgh schools one that encourages and supports attendance? What changes might make the schools places where students want to be? What are the barriers to improved attendance and how might they be removed?

For questions about this report or more information about the National Center for Schools and Communities at Fordham University contact us at
33 West 60th Street Second Floor, New York NY 10023,
212-636-6699 or visit our web site at www.NCSCatFordham.org


Appendices

Table 1. Significant relationships between attendance and math and reading levels in Pittsburgh's public schools
Attendance
Elementary Pearson's r(n)
Middle Pearson's r (n)
High Pearson's r (n)
Percentage of students scoring in top level (4) of 5th grade math test
.398*
(59)
Percentage of students scoring in top level (4) of 5th grade reading test
.476**
(59)
Percentage of students scoring in level 3 of 5th grade reading test
.303*
(59)
Percentage of middle school students scoring in top level (4) of 8th grade math test
.619**
(20)
Percentage of middle school students scoring in level 3 of 8th grade math test
.728**
(20)
Percentage of middle school students scoring in top level (4) of 8th grade reading test
.590**
(20)
Percentage of middle school students scoring in level 3 of 8th grade reading test
.748**
(20)
Percentage of students scoring in top level (4) of 11th grade math test
.761**
(11)
Percentage of students scoring in level 3 of 11th grade math test
.741**
(11)
Percentage of students scoring in top level (4) of 11th grade reading test
.730**
(11)
Percentage of students scoring in level 3 of 11th grade reading test
.807**
(11)

*p< .05, **p< .01


Table 2. Significant relationships between attendance and math and reading scale scores in Pittsburgh's public schools
Attendance
Elementary Pearson's r(n)
Middle Pearson's r (n)
High Pearson's r (n)
Fifth grade math scale score
.353**
(59)
Fifth grade reading scale score
454**
(59)
Eight grade math scale score in middle schools
.782**
(20)
Eight grade reading scale score in middle schools
.786**
(20)
Eleventh grade math scale score
.861**
(11)
Eleventh grade reading scale score
.844**
(11)


*p< .05, **p< .01


Table 3. Significant relationships between teacher experience and qualifications and academic performance indicators

Academic Performance Indicators
Average years of teaching experience Pearson's r
(n)
Percent of teachers with a master's degree Pearson's r
(n)
Average test scores on fifth grade math test
.319*
(56)
.363**
(56)
Average test scores on fifth grade reading test
.470**
(56)
.384**
(56)
Average test scores on eleventh grade math test
.743**
(11)
Average test scores on eleventh grade reading test
.843**
(11)
Percentage of students scoring in level 1 of 5th grade math test
-.358**
(56)
-.309*
(56)
Percentage of students scoring in level 3 of 5th grade math test
.406**
(56)
Percentage of students scoring in level 1 of 5th grade reading test
-.427**
(56)
-.365**
(56)
Percentage of students scoring in level 4 of 5th grade reading test
.449**
(56)
Percentage of middle school students scoring in level 4 of 8th grade math test
.470*
(18)
Percentage of middle school students scoring in level 2 of 8th grade reading test
-.485*
(18)
Percentage of middle school students scoring in level 4 of 8th grade reading test
.511*
(18)
Percentage of students scoring in level 1 of 11th grade math test
-.777**
(11)
Percentage of students scoring in level 2 of 11th grade math test
.764*
(11)
Percentage of students scoring in level 1 of 11th grade reading test
-.856**
(11)
Percentage of students scoring in level 3 of 11th grade reading test
.831**
(11)
Percentage of students scoring in level 4 of 11th grade reading test
.628**
(11)

*p< .05, **p< .01


Table 4. Significant relationships between programmatic resources, school performance indicators and teacher qualifications

School Performance Indicators
Total number of academic programs
Pearson's r
(n)
Total number of supportive programs
Pearson's r
(n)
Elementary enrichment program
Pearson's r
(n)
Average test scores on fifth grade math test
419**
(59)
Average test scores on fifth grade reading test
.259*
(59)
Percentage of students scoring in level 4 of 5th grade math test
.325*
(59)
Percentage of students scoring in level 3 of 5th grade math test
.405**
(59)
Percentage of teachers in elementary school with a master's degrees
.305*
(56)
Middle school attendance
.519*
(20)

*p< .05, **p< .01


Table 5. Significant relationships between student demographics and school resources

 

School Resources
Percent of low income students
Pearson's r
(n)
Percent African American
Pearson's r
(n)
Percentage of teachers with a master's degree in all schools
-.359**
(87)
-.271*
(88)
Percentage of teachers with a master's degree in elementary schools
-.309*
(56)
-.338*
(56)
Average years of teaching experience in elementary schools
-.349**
(56)
-.389**
(56)
Percentage of teachers with a master's degree in middle schools
-.530*
(18)
Average years of teaching experience in middle schools
-.555*
(18)
Average years of teaching experience in high schools
-.737**
(11)
Total number of academic programs in elementary schools
-.412**
(59)

*p< .05, **p< .01


Table 6a. Significant association between enrichment program and school characteristics

School Characteristic Elementary Enrichment Program
Schools with program (n=41) School without program (n=18)
Mean sd Mean sd t
Percent of low-income students in elementary school 65.08% 18.36 76.57% 13.78 t =-2.373*


Table 6b. Significant association between foreign language program and school characteristics

School Characteristic Elementary Certified Foreign Language Program
Schools with program
(n=16)
Schools without program
(n=43)
Mean sd Mean sd t
Percent of low income students in elementary schools 54.73% 16.34 73.73% 15.54 t =-4.119**

*p< .05, **p< .01


Elementary School Programs 2000-2001 (Grades K-6)

Academic Programs/Opportunities/Initiatives Supporting Programs/Opportunities/Initiatives
Full-day kindergarten Before school programs/clubs
Half-day kindergarten After school programs/clubs
Art instruction with certified art instructors Intramural sports
Music instruction with certified music instructors Band/orchestra
Acceleration programs Chorus
Enrichment programs Parent involvement programs/organizations
Tutorial or extra help programs Business partnerships
Magnet and/or academy programs Even start
Environmental education center Community service programs/opportunities
Foreign language instruction with certified foreign language instructor On-site lunch service
Physical education instruction with certified physical education instructor On-site breakfast service
Independent study courses Head Start
Educational field trips
Distance learning
School to work activities
Totals
15
12


Middle/Junior High School Programs 2000-2001 (Grades 6-9)

Academic Programs/Opportunities/Initiatives Supporting Programs/Opportunities/Initiatives
Required art courses Before school programs/clubs
Required music courses After school programs/clubs
Acceleration programs Intramural sports
Enrichment programs Band/orchestra
Tutorial or extra help programs Chorus
Magnet and/or academy programs Theater/arts activities or productions
Environmental education center Parent involvement programs/organizations
Foreign language courses Business partnerships
Required physical education courses Community service programs/opportunities
Distance learning On-site lunch service
Industrial arts/technology education On-site breakfast service
Ninth-grade vocational education program(s)
Career exploration/career resource center
School to work activities
Consumer and homemaking education
Tech prep
Work-based learning
Independent study courses
Honors programs/courses: Math, Science, English, Social Science, Arts, Other
Totals
26
11


Senior High School Programs 2000-2001 (Grades 9-12)

Academic Programs/Opportunities/Initiatives Supporting Programs/Opportunities/Initiatives
Art course cluster or major Before school programs/clubs
Music course cluster or major After school programs/clubs
Acceleration programs Intramural sports
Enrichment programs Interscholastic sports
Tutorial or extra help programs Band/orchestra
Magnet and/or academy programs Chorus
Environmental education center Theater/arts activities or productions
Foreign language courses (Non-traditional) Community service programs/opportunities
Foreign language courses (Level 5 and above) Parent involvement programs/organizations
Required physical education courses Business partnerships
Distance learning Work study
Independent study courses On-site lunch service
School to work activities On-site breakfast service
Consumer and homemaking education
Tech prep
Work-based learning
High-Schools-That-Work initiative
Higher education courses offerings
Career exploration/career resource center
Industrial arts/technology education
Driver education
Honors programs/courses: Math, Science, English, Social Science, Arts, Other
Totals
28
13