A Comparison Of The Performance Of The Milwaukee Public Schools And School Systems In Selected Other Cities

by
Gerald W. Bracey

Center for Education Research, Analysis, and Innovation
School of Education
University of Wisconsin-Milwaukee
PO Box 413
Milwaukee WI 53201
414-229-2716

January 2000

CERAI-00-03

AComparison Of The Performance Of The Milwaukee Public Schools And SchoolSystems In Selected Other Cities

GeraldW. Bracey

January2000

Executive Summary

This report examines the city of Milwaukee, Wisconsin, on a number ofdemographic variables and on test scores, and compares the data from Milwaukeeto those from 10 other cities.

Among the findings are that of the 11 cities, Milwaukee ranks at or near thebottom in terms of adult education levels (10^th), child povertyrates (9^th) and per capita income (11^th). In its schools,Milwaukee has larger pupil/teacher ratios than most cities (17.6 to 1). Itspends somewhat more on schools than some other cities, but it ranks first infocusing its fiscal resources on instruction and the support of instruction.

Milwaukee has a lower dropout rate and higher test scores than mostcomparison cities, particularly in relation to those cities that most closelymatch Milwaukee’s demographics.

The correlation between parental educational level and test scores is wellknown. The impact of poverty on achievement is also well documented. One of thebest known such studies – and one that was unique in the way it looked at thesubject – examined the performance of students in high- and low-povertyschools.¹ The study was conducted for the USDepartment of Education during the Bush administration by the consulting firmof Abt Associates in Cambridge, Mass. In this study , "high-poverty"schools were defined by having 76% or more of the students eligible for free orreduced-price lunch. "Low-poverty" schools had zero to 20%eligibility. Researchers first divided students into categories based on thetypical letter grades they took home on report cards. They then looked to seehow the various categories performed on standardized tests of reading andmathematics. Students in low-poverty schools who received "A’s"scored at better than the 80^th percentile on both tests. Students inhigh poverty schools who received "A’s" scored higher than studentsin high-poverty schools who received lower grades, but on neither test did theiraverage score reach the 40^th percentile. In sum, the studydemonstrated that poverty greatly depresses academic achievement.

The most recent scores available from Milwaukee that can be compared withschool districts in other states are from the 1993-’94 school year. After thatyear, Milwaukee stopped using norm-referenced tests that were validated againstnational samples; consequently, the district’s more recent tests cannot beaccurately compared with districts in other states. Additionally, for that year,the available test scores were for Grade 5 only. In the 1993-’94 tests, fifthgraders performed at the 38^th percentile in reading and the 47^thpercentile in math. While this performance is certainly not one to inducecomplacency, in view of the high poverty and low education levels of Milwaukeeadults, and in view of the test scores of similar cities, the performance issurprisingly high.

We should note that the national norm for a standardized test score is the50^th percentile. While Milwaukee students score below this nationalnorm in reading, they are extremely close to it in math. While there is clearlyroom for improvement, it would be misleading to characterize Milwaukee as a"failing" system.

Introduction

This report examines the city of Milwaukee, Wisconsin, on a number ofdemographic and school variables in the elementary grades, and compares thedata from Milwaukee to those from 10 other cities, chosen because their sizeand ethnic makeup make them roughly comparable to Milwaukee.

The cities chosen for comparison were selected because two key variables,size and ethnic makeup, gave them a rough similarity to Milwaukee. Thesedistricts are Boston, Mass.; Buffalo, N.Y.; Cincinnati and Columbus, Ohio; FortWorth, Texas; Indianapolis, Ind.; Kansas City, Mo.; Nashville, Tenn.;Sacramento, Calif.; and Jacksonville, Fla.

All of the demographic data for this report were obtained from the SchoolDistrict Data Book, compiled for the US Department of Education and obtainablefrom the National Center on Education Statistics.²The Data Book provides a rich set of variables from the 1990 census and the USDepartment of Education’s Common Core of Data for the same year. The Data Bookcontains the most comprehensive data currently available. Test score data wereobtained either from web sites or by contacting the assessment offices of thevarious districts.

Characteristics of Milwaukee

The cities under review ranged widely on the demographic variables examinedfor this report, but on many measures, Milwaukee appeared at or near thebottom.

Milwaukee has the second-lowest adult educational attainment of the citiesstudied. It ties with Indianapolis in having the smallest percentage of peoplewith at least a bachelor’s degree, 14%, and has a relatively high proportion ofpeople who did not complete high school, 27%. Indianapolis has the highestproportion of people without diplomas, 33%. At the other end of the spectrum,Boston has the highest proportion with college degrees, 30%.

Milwaukee has the third-highest poverty rate, 22%, and second-highest childpoverty rate, 37%. Buffalo has the highest poverty and child poverty rates, 26%and 38%, respectively. Jacksonville ties Nashville for the lowest povertyrates, 13%, and has the lowest child poverty rate, 18%.

Milwaukee is in a virtual tie with Indianapolis for lowest per capitaincome, $11,106, while Boston has the highest, $15,581. Milwaukee is most likeKansas City in its ethnic makeup.

The pupil/teacher ratio in Milwaukee is exceeded slightly by two otherdistricts, Jacksonville and Fort Worth, and exceeded greatly by one,Sacramento. Milwaukee’s ratio is 17.6 to 1, while Sacramento’s is 23.3 to 1. Itshould be kept in mind that pupil/teacher ratio is not the same as class size. Somecity districts with low pupil/teacher ratios have very large classes.Pupil/teacher ratios typically include all employees with teachingcertificates, whether or not they have classroom duties. It also includesteachers in special education and limited English proficiency classes, whichare typically small. Therefore, pupil/teacher ratio should not be confused withclass size.

Three districts spend more than Milwaukee on pupils and seven districtsspend less. How those monies are allotted however, varies considerably.Milwaukee spends 58% of its funds on instruction, about average for the citiesexamined, and spends 35% of its funds on support services – making it firstamong all the cities in this category. "Instruction" includes teachersalaries, textbooks, and instructional materials. "Support services"includes special education, counseling, instruction to students with limitedEnglish proficiency, and related costs. Milwaukee’s allocation of funds makesit the top spender on instruction and support combined, with 93% of all fundsallocated to these two categories. Nashville spends 70% of its funds oninstruction, while Kansas City spends only 34%. Nashville spends only 18% ofits funds for support services, while Kansas City and Columbus both spend 34%.Kansas City has the lowest combined proportion, 68%.

No discussion of urban schools would be complete without looking at dropoutrates, but the way those rates are normally presented creates challenges for areport such as this one focusing primarily on elementary students and seekingto make comparisons across districts. Dropout rates are normally compiled interms of raw numbers – which cannot be accurately compared among districts ofdifferent sizes – or in terms of percentage of high school students. Forpurposes of this report, however, a dropout rate was developed for eachdistrict, dividing the number of dropouts by the district’s total K-12enrollment.

Using this measure, Milwaukee in 1990 had the third lowest dropout rate,5.1%, while Indianapolis has the highest, 8.8%, and Buffalo the lowest, 4.6%.

All of the districts in 1990 had at least 30% white enrollment exceptBoston, with 25%. Maximum white enrollment proportion is 60% for Jacksonvilleand Nashville. All districts had at least 33% African-American enrollmentexcept Sacramento which has 20%. Maximum African-American enrollments are inCincinnati (59%), Kansas City (57%) and Milwaukee (55%). Hispanic enrollmentsvary from not registering in Columbus to 29% in Forth Worth. Asian enrollmentsvary from not registering in Fort Worth, Indianapolis or Columbus, to 22% inSacramento.

Test Scores

Five different tests are represented in the data. They are: the Iowa Testsof Basic Skills (Milwaukee, Boston, Kansas City), the Metropolitan AchievementTest (Forth Worth), Terra Nova (Indianapolis, Jacksonville), the ComprehensiveTests of Basic Skills (Jacksonville, Nashville) and the Stanford-9 (Cincinnati,Boston, Sacramento). Buffalo tests children only with tests specific to thestate of New York and therefore cannot be accurately compared with the otherdistricts. The metric used for reporting all tests is the national percentilerank, except for Cincinnati. For that district, all that was available was thepercent of students above the national norm. However, in instances where bothNPRs and percent above the national norm were available, they were identical ornearly identical. For instance, the two statistics for Forth Worth are as shownin Table 1:

Table 1: National Percentile Rank and % above norm/Fort Worth Schools

	National Percentile Rank	% Above Norm
Grade 3
Reading	49	49
Math	45	45
Grade 4
Reading	42	41
Math	50	53
Grade 5
Reading	45	45
Math	55	57

For the purposes of the comparisons in this report, therefore, the twometrics appear to be interchangeable.

Finally, while comparisons of the same metric on different tests areimprecise, they are of at least limited value to the extent that the norms forthe various tests come from valid representative samples of students and to theextent that sound psychometric procedures are used.

Standardized achievement tests are most commonly administered in grades 3,5, 8 and 11, but the actual grades in which tests are given vary from districtto district. For this report, test scores were compiled for all grades in whichtests were given in each district in the study. For Milwaukee scores werechosen for the 1993-’94 school year, because that is the last year in whichMilwaukee administered norm-referenced achievement tests that could be comparedwith test results from other states. For Milwaukee that year, the onlycomparable test results available were for grade 5 students.³

Scores from other cities were primarily from 1997, 1998 or 1999. In theelementary years, there is no great, systematic difference among scores fromdifferent grade levels, particularly when the metric used is percentile rank.For that reason, meaningful comparisons can be made between scores administeredin different school districts in different years, and between percentile ranksfor tests administered in different years.

Milwaukee’s test scores are shown in Table 2, below, followed by scores inother cities in Tables 3 through 11.

Table 2

Milwaukee – ITBS
Grade	Reading	Math
Grade 5	38	47

Other Cities:

Table 3

Boston – Stanford 9
Grade	Reading	Math
Grade 4	42	48
Grade 5	47	50
Grade 6	44	45

Table 4

Cincinnati – Stanford 9
Grade	Reading	Math
Grade 3	33	37
Grade 4	35	34
Grade 5	37	32

Table 5

Columbus – Metropolitan Achievement Tests
Grade	Reading	Math
Grade 3	32	36
Grade 5	40	37

Table 6

Fort Worth – Stanford 9
Grade	Reading	Math
Grade 3	49	45
Grade 4	42	50
Grade 5	45	55

Table 7

Indianapolis – Terra Nova
Grade	Reading	Math
Grade 3	38	45
Grade 6	29	35

Table 8

Jacksonville (Duval County), Fla. – Terra Nova (Grade 3); CTBS/4 (Grades 4, 5, and 6)
Grade	Reading	Math
Grade 3	53	50
Grade 4	52	58
Grade 5	50	57
Grade 6	47	51

Table 9

Kansas City – ITBS
Grade	Reading	Math
Grade 3	29	32
Grade 5	32	35
Grade 6	30	30

Table 10

Nashville – CTBS
Grade	Reading	Math
Grade 3	49	49
Grade 4	47	47
Grade 5	45	37
Grade 6	41	40

Table 11

Sacramento – Stanford 9
Grade	Reading	Math
Grade 3	37	47
Grade 4	38	39
Grade 5	38	43

One thing that is not known is the policy on student exclusion from testing.These vary from district to district and can have an impact on results.

In virtually all instances, the performance on mathematics exceeds that inreading. At these grade levels increased performance in mathematics canprobably attained through extended drill-and-practice exercises, whileincreases in reading cannot.

Milwaukee is among the few districts that supplied any ethnic orprogrammatic differentiations. In some cases, ethnic differences were shown forother parts of the testing program, but not the norm-referenced component; inothers it appears that such categories are not used.

Table 12

Milwaukee: Grade 5
	Reading	Math
White	64	72
African-American	26	35
Hispanic	39	51
Asian	43	68
Native	41	60
Free Lunch	30	41
Non-Free Lunch	63	68

The city that Milwaukee most closely resembles demographically, Kansas City,has the lowest test scores of all 11 cities in the study. We should keep in mind,though, that the Kansas City school system has been embroiled in controversyfor a number of years now. One report found that during one 9-year period, thedistrict went through 10 superintendents.⁴ During the same period, many candidates withdrew their namesafter they had actually met with the school board. Still, it is worth notingthat one place where the two cities differ dramatically is in how funds areallocated. While the two have similar proportions spent on support services,Milwaukee spends a substantially higher proportion on instruction.

Indianapolis is also demographically similar to Milwaukee. The Grade 3 testscores for Indianapolis are similar to Milwaukee’s Grade 5 scores; Grade 6scores in Indianapolis are lower than the Milwaukee Grade 5 scores. Milwaukeeand Indianapolis spend money in comparable ways.

The one city that consistently outperforms Milwaukee on test results,Jacksonville, Fla., has a better-educated population, a different ethnicmakeup, more wealth, and the lowest child poverty rate of all 11 cities.

In a recent Educational Testing Service study on parental satisfaction,Milwaukee finished slightly below average among the 55 cities studied.⁵ The study reported the percentage of parents who weredissatisfied and the percentage of parents so dissatisfied they wanted to move.

For the eight cities from the ETS study that are also in this study, Table13 shows the percentage of parents dissatisfied with the schools, and eachcity’s rank in the overall ETS study.

Table 13

City	% dissatisfied with schools	Rank among 55 cities in study
Cincinnati	34.2	3^rd
Boston	19.5	15^th
Buffalo	18.9	16^th
Milwaukee	17.4	21^st
Kansas City	16.9	24^th
Columbus	16.6	26^th
Indianapolis	14.5	35^th
Sacramento	13.9	37th

Table 14 displays the percentage of parents in each city who were sodissatisfied that they wanted to move out of the district, according to the ETSstudy.

Table 14

City	% saying they want to move
Cincinnati	13.7
Buffalo	12.7
Columbus	9.3
Kansas City	9.2
Boston	8.7
Sacramento	8.2
Milwaukee	7.5
Indianapolis	7.2

Table 13 shows that Milwaukee is in the middle of the pack in terms ofoverall dissatisfaction. But as Table 14 shows, only Indianapolis has a lowerpercentage of people who are angry enough to want to move. Given thehigh-profile national and local criticisms of the Milwaukee public schools,this figure seem remarkable; one would expect many more parents to express thedesire to move out of the district. One might speculate that the presence of aschool voucher program in Milwaukee dampens some of the desire to move, asparents would reason that they have access to alternative schools withoutmoving. Because the voucher program is available only to a limited number offamilies, however, this explanation seems unlikely.

Discussion

The data examined here is less precise than would be ideal, both becausethey come from a variety of sources that are not always precisely comparable,and because of the age of some of the information. Even so, they are the bestdata – in some cases the only data –available to make the sort of comparisonsbeing attempted.

Despite a degree of imprecision, the data are striking enough that twogeneralizations are warranted. First, Milwaukee students score better onstandardized tests than Milwaukee’s demographic characteristics would lead oneto predict. In those communities with characteristics most similar toMilwaukee’s, test scores are lower. Second, in communities with one or twodemographic variables that are more favorable than Milwaukee’s, test scores aresimilar. Only cities with generally more favorable demographics showconsistently higher test scores than Milwaukee.

One reasonable-sounding objection to the comparison of older Milwaukee datawith newer data from other cities arises if one speculates that Milwaukee’stest scores have fallen in the intervening years. Recent reports from theMilwaukee Public Schools, however, indicate just the opposite: The most recentreports from the Milwaukee Public School Administration show improvements inboth test scores and dropout rates (as the district defines the term, not asdefined in this report) in the 1998-’99 school year.⁶Therefore, it seems likely that using an earlier set of statistics for thedistrict than for other cities in the comparison – while providing morereliable comparisons for reasons already discussed – may actually understatethe district’s performance compared with its out-of-state peers.

Milwaukee is high in poverty, particularly high in child poverty, low inincome and low in adult educational level in comparison to the other cities.All of these variables mean that there are fewer resources in the home toassist in the educational development of the children. People with littleeducation cannot provide direction; people with little money cannot provide,for example, such things as educational toys.

Although poverty is sometimes dismissed as an "excuse" for lowachievement, it is more appropriately viewed as a condition that, like gravity,affects virtually everything about a child’s life. Poor pregnant women arelikely to get prenatal care late if at all. Poor children are about twice aslikely as middle-class children to have physical and mental disabilities andare seven times more likely to be abused.⁷

The area indicating the highest need for attention is the achievement ofMilwaukee’s African-American students. Their scores are considerably lower thanthe national norm and substantially lower than for any other minority group.Although the reading score is the more depressed (29^th percentile),the math score (35^th percentile) is a cause of real concern in theface of Milwaukee’s algebra-for-all requirement. Attention must be paid toincreasing the achievement of Milwaukee’s African-American students. One place tostart might be to analyze the funds that currently go into the"support" category to see if some of them might be better re-deployedin the "instruction" arena directly. However, at noted earlier,Milwaukee does a relatively good job of focusing money on instruction.

Overall, though, given Milwaukee’s child poverty rate, income levels, andadult educational levels, the test scores found for Milwaukee are higher thanone would expect. The scores compare especially well with those cities thatmost closely resemble Milwaukee on these variables.

These findings also suggest that the work of improving the educationalprospects for Milwaukee schoolchildren does not belong to MPS alone. The schooldistrict is in some respects already doing better than expected; what it mayneed most is appreciation for what it has done, a challenge to continue toimprove its performance, and perhaps most of all, for other institutions in thecommunity to do their part in alleviating the conditions of poverty, poor adulteducational attainment, and racial isolation that afflict the environment inwhich the schools operate.

ENDNOTES:

1. Reinventing Chapter 1:The Current Chapter 1 Program and New Directions. Washington, DC: U. S.Department of Education, Office of Policy and Planning, February, 1993.

2. The School District DataBook, 1989-1990. Washington, DC: National Center for Education Statistics,1990. Available: http://govinfo.kerr.orst.edu/sddb-stateis.html. [1 February, 2000]

3. Personal Communication,Milwaukee Public Schools division of research, October 1999.

4. Paul Ciotti, Money andSchool Performance. Washington, DC: Cato Institute, March 1998.

5. Anthony P. Carnevale andDonna M. Desrochers, School Satisfaction: A Statistical Profile of Citiesand Suburbs. Princeton, NJ: Educational Testing Service, 1999.

6. Joe Williams, "MPSStudents Making Gains," Milwaukee Journal Sentinel, Jan. 26, 2000.

7. Helping Families Work.Columbus, Ohio: Ohio Children’s Defense Fund, 1994.