Wisconsin Policy Research Institute Report:

The Achievement Gap in Milwaukee Public Schools:
Outcomes by Gender, Race, and Income Level

By Sammis White, Ph.D.

Table of Contents:

I.      Executive Summary
II.     Introduction
III.    Research
IV.    Data
V.     Methodology
VI.    Analysis
VII.   Policy Implications and Interventions
VIII.  Recommendations
IX.    Conclusion

  EXECUTIVE SUMMARY

Despite a host of initiatives the Milwaukee Public Schools (MPS) have had very modest and irregular success in raising levels of student achievement in recent years. The results are not sufficient to ensure economic health for the individual students or the regional economy. These are among the numerous compelling reasons why achievement levels must rise.

To raise student achievement, we first need a better understanding of what factors may be most influential in the low scores that have been commonplace in MPS. A factor that may be playing a role is the failure to teach in ways that help the majority of males achieve at higher levels. Increasingly, evidence suggests urban school systems fail males, especially minority males. According to one source, African-American males in MPS, for example, have a graduation rate of about 31%, and Hispanic males have an estimated graduation rate of 36%. The white male rate of graduation from MPS is estimated to be 66%, suggesting something is out of kilter for minority males, not to mention the system as a whole, one that largely serves a low-income population.

We must also note, however, that minority females, while doing better than comparable males, have not achieved at close to white female levels in the district. The MPS African-American female graduation rate is estimated at 46%, Hispanic female, 50%, and white female, 75%, in a state where the average overall graduation rate is 88.8%. Graduation is just one measure, albeit a critical one, of achievement. Thus, the question is whether the problem of low achievement is one of gender, race, or income.

This report examines the differing levels of achievement of two genders, three different racial groups, and four different levels of income among two classes of students followed over seven or eight years to learn what factors appear to have the greatest impacts on student achievement in reading and math. The analysis is done of cohorts of Milwaukee Public School students.

Among the findings that appear are the following stark messages:

• On reading, MPS male students of all incomes and races combined, on average, score consistently below similar females. The gap increases substantially as students move to higher grades.

• On math, MPS male students of all incomes and races combined tend to score, on average, just marginally below females, and the gaps do not change much as students move to higher grades.

• Within the same gender there are very large gaps in average reading and math scores between students with the lowest incomes (eligible for free lunches) and middle-income students (those who did not apply for lunch support), starting in the earliest years of testing. The sizes of these gaps grow markedly among both males and females, as students move to higher grades.

• Large gaps in average reading achievement scores exist between minority and white MPS students (all incomes combined). The scale of these gaps grows as students move through the grades.

• Among MPS females the average Hispanic/white reading gap, on average, is relatively modest most years, but the average African-American/white gap starts large at lower grades and doubles by tenth grade. This latter gap most years is roughly estimated at more than a year of progress.

• Among MPS males the average Hispanic/white reading gap, on average, is larger than the female gap, but it is in the range of about a half-year’s progress until tenth grade, at which point it increases. The average African-American/white gap starts early and remains large, increasing dramatically in tenth grade. Most years the gap is roughly estimated at more than a year of progress.

Despite modest average differences between males and females in math, there are large differences in average math achievement scores between minority and white MPS female and male students (all incomes combined). The scale of these gaps grows as students move through the grades.

• Among females the average Hispanic/white math score gap begins at a modest level and doubles within a few years. The average African-American/white gap begins large and grows substantially over the grades 4th-10th. Most years the gap is very roughly more than a year of progress.

 

• Among males the average Hispanic/white math score gap is large (over half a year) at fourth grade and basically doubles by high school. The average African-American/white gap begins large at twice the Hispanic/white gap and almost doubles over grades 4th-10th. Most years the Black/white gap is very roughly more than a year of progress.
  

• When all three characteristics are combined and the consistently lowest-scoring group of students’ (on average, African-American males) achievement levels are compared with the group that is consistently the highest-scoring group, white females, by grade and subject, extremely large differences are commonplace.
  

• For the Class of 2011 the difference in average reading achievement between these two groups is 49 points at fourth grade, rising to 70 points in seventh grade. This can loosely be estimated to be a three-year difference in reading levels, on average, by seventh grade.
  

• For the Class of 2011 the difference in average math achievement between these two groups is 30 points at fourth grade, rising to 61 points in seventh grade. This can loosely be estimated to be a two- to three-year difference in math levels, on average, by seventh grade.

If measured against the average student in the state of Wisconsin at 8th grade, the average African-American male in MPS is approximately two years behind in math and almost two years behind in reading.[i]1 Those are huge gaps that should be totally unacceptable to the citizens of the state.

Hispanics have also been achieving at lower than white levels, on average. They should be given every opportunity to achieve at least at middle-income white levels. So programs should not be targeted just at African-Americans or African-American males.

Furthermore, the gap is not just a minority gap: on average, low-income white males are often a year or more behind middle-income white males in reading and math. That indicates a very large need for instituting approaches that are targeted at low-income children, regardless of race. A high proportion of these children could use the attention.

These learning achievement differences must be addressed both for the sake of the individuals and for the sake of the regional economy. The Milwaukee economy is growing increasingly reliant on a minority workforce. If that workforce is largely undereducated, then the economy will not be able to compete globally, and incomes of all residents, not just the undereducated, will suffer.

The learning gaps must be addressed, beginning now. Steps must be taken both inside and outside MPS to raise student achievement and high school graduation rates for all. To begin the move in that direction, some suggestions are offered. Each has some research support, but several would benefit from additional application and evaluation in the Milwaukee setting. The most important step is taking more concerted efforts to raise levels of student achievement.

There are several steps that are very likely to make success easier to achieve in K-12 education. Some of these steps are harder to achieve than the others, although the reader may have trouble deciding which of the five mentioned is really harder—since all will be opposed to at least some degree. The five recommendations are:

•         Increase parent involvement in their child’s education.

•         Inject accountability in the governor, legislature and local educators for student outcomes.

•         Replicate lessons from successful schools.

•         Better prepare children for school before they reach Kindergarten.

•         Embed body-movement exercises in everyday classes in all schools and preschools.

The author would like to acknowledge the assistance of Bruce Thompson in the development of the data for this analysis. The author is solely responsible for the analysis and policy recommendations.

INTRODUCTION

In the last year numerous headlines have appeared in the news media that suggest a large gender problem exists in our schools. The New York Times said: “Boys are No Match for Girls in Completing High School.”[i] Another from the Times: Dire Problems for Young Black Men, Several New Academic Studies Warn.”[ii] Locally, the Journal Sentinel featured a story that “Boys learn differently from girls, studies say.”[iii] And last July in the results of a national study Wisconsin fared worst in the nation in what was termed the state Education Inequality Index, the difference between graduation rates of black males and white males. The gap in Wisconsin was 47 points, the difference between a graduation rate of 38% for black males and 84% for white males across the state.[iv]

There is also a growing literature that bemoans the treatment boys are receiving in education today.[v] The claim is that males learn differently and that those differences work against them in schools that teach to the girls’ way of learning. They cite evidence of higher grades, higher high school graduation rates of girls, higher rates of college attendance among young women, and higher grades in college. Boys, these critics complain, are getting the short end of the stick.

On the other hand, another group claims that the issue is that girls are finally being given the education that they deserve and that both sexes are doing better overall.[vi] One only need to note that the majority of college students today are women to know that girls are finally competing well with boys, at least on many indicators.

But even those who argue that females are only beginning to equal males realize that certain males—minority and low-income—are not doing well. These defenders of girls’ achievements admit that “academic performance for minority boys is often shockingly low.”[vii] This is certainly the case in Milwaukee where minority test scores and graduation rates are significantly behind those of white students in the district and even further behind those of white students in the rest of the state.

Milwaukee Public Schools’ (MPS) test scores have not been rising for several years, despite a host of initiatives (Figure 1). The district has seen an increase in the proportion of low-income students, so holding scores steady may be an accomplishment, modest though it may be. But stability in scores is not enough. There are numerous compelling reasons why achievement levels must rise. A recent report on MPS reveals that outside reviewers see MPS as far too complacent with its non-gains.[viii] The reviewers insist that more pressure be placed on MPS for gains in student achievement.

Pressure alone is not sufficient. We need a better understanding of what factors may be most influential in the low scores that have been commonplace in MPS. The new report cites the decentralization of authority as a contributor to no steady gains. MPS administration is now trying to reverse that (re-centralize), so that it can have a greater role in curriculum and budget decisions, among other points. But few view management style as the determining factor in educational outcomes.

A factor, however, that may be playing a role is the failure to teach in ways that helps the majority of males achieve at higher levels. According to one source, African-American males in MPS, for example, have a graduation rate of about 31%, and Hispanic males have an estimated graduation rate of 36%.[ix] The white male rate of graduation from MPS is estimated to be 66%, suggesting something is out of kilter for minority males, not to mention the system as a whole. What must also be noted is that the minority females, while doing better than comparable males, have not achieved at close to white female levels in the district. The African-American female graduation rate is estimated at 46%, Hispanic female, 50%, and white female, 75%. This is just one measure, albeit a critical one, of achievement.

Do these differences among these groups start at young ages or do they develop over time, perhaps as peer pressure gets stronger? Do the differences vary by subject area or are boys behind girls in levels of achievement on every subject? And do Hispanic males and females more closely follow the patterns of achievement for whites or for African-Americans? These are important questions to answer, because the answers can lead to more appropriate curricula and teaching methods.

If gender differences are not an issue, then we can look for other factors that might make better targets for interventions. Given the media coverage, we look first to the gender differences.

RESEARCH

The popular press has been quick to pick up on the proposition that males, especially African-American males, are falling further and further behind not only white males but African-American females. Business Week ran an article entitled “The new gender gap: From kindergarten to grad school, boys are becoming the second sex.”[x] The article made the case that boys are not doing as well in school achievement, school graduation, or college entrance or graduation. And they make the case that this is a problem, not only economically but socially. USA Today ran an opinion piece entitled: “Pay closer attention; Boys are struggling academically.”[xi] It talks of the reforms that have helped girls do better and that several of the reforms have made achievement more difficult for boys. The impression given is that boys are struggling.

Despite these media articles on males, the evidence is not so clear. A review of trends in the educational equity of girls and women for 2004 reveals that “females have done much better than males in reading and writing, but have generally, though not always, lagged behind in science and mathematics.” The report also states that “females in grades 4, 8, and 12 have consistently outperformed males in reading . . . and also outperformed their male peers in writing in 1998 and 2002.”[xii]

The same study revealed a somewhat different finding with regard to math. In fact, “[a]lthough there is a common perception that males consistently outperform females in mathematics, National Assessment of Educational Progress (NAEP) mathematics scores have not shown this. In mathematics, the gap between average scale scores has been quite small and fluctuated only slightly between 1990 and 2003.”[xiii] Another study, one that examined 111 studies on male and female abilities concluded that “most of the [studies] suggest that men’s and women’s abilities for math and science have a genetic basis in cognitive systems that emerge in early childhood but give men and women on the whole equal aptitude for math and science.”[xiv] The author goes on to report that boy and girl infants were found to perform equally well as young as six months on such tasks as addition and subtraction.

Surprisingly, given the results of the NAEP test scores on reading and writing, Hyde and a colleague reported that data from 165 studies revealed a female superiority so slight as to be meaningless, despite previous assertions that “girls are better verbally.”[xv] Hyde and two colleagues examined math performance and concluded that there was little support for saying boys are better at math. They instead concluded that social and cultural factors influence perceived or actual performance differences.[xvi]

Reinforcing the conclusion that social and cultural factors are very influential were the results from the two international tests. One is from the Trends in Mathematics and Science Study (TIMSS) for students in grades 4 and 8. The second is the Program for International Student Assessment (PISA) for students at age fifteen. The American Institutes for Research (2003) noted that boys in the United States consistently outperform girls in all three assessments on math. The differences are small (less than a tenth of a standard deviation). But the most intriguing finding is that the U.S. and Italy are the only countries out of twelve compared in which boys consistently outperform girls on all three assessments.[xvii]

The case for culture and social factors being influential may well play a role in what several studies see as a greater gap between males and females—that between African-American males and females. According to a 2004 report (The Schott Foundation for Public Education) the widest gap separates African-American males from other sub-groups of students, including Black females.[xviii] The study notes that several school districts have the lowest Black male graduation rates in the country; these include: Cincinnati and Cleveland, Ohio, 19%; Chatham County, Georgia, 21%, Rochester, New York, Milwaukee, Wisconsin, and Pinellas County, Florida, 24%.[xix] Although these numbers likely overstate the problem, it is clear that the African-American, male graduation rates are very low in Milwaukee and elsewhere.

It is this phenomenon that needs greater exploration. Why is it that African-American males have such low graduation rates? Does this pattern of modest achievement start at some early date? If it does, can interventions be identified that can help to reduce the failure and desire to leave high school before it can be completed? These are critical questions. Again, the popular media has picked up on this and the implications. The USA Today editorial referred to above states that for African-American men the gender gap is widening at an alarming rate in terms of high school completion, college enrollment, and college degrees. The 2000 US Census pointed out that 35% of African-Americans enrolled in college were men.

The primary quest for this report is to learn if and when African American males in Milwaukee begin to fall behind African-American females in terms of reading and math. The report also seeks to learn if the pattern is different among Hispanics and whites in the Milwaukee Public Schools. If the African-American males drop behind at an early age, the question then becomes what sorts of interventions might be appropriate to try to raise average African-American male scores to at least equivalent to those of average African-American females and even better would be to make them equivalent to white males and females. 

DATA

To explore the pattern of student achievement within the Milwaukee Public Schools, we employed the data resources of MPS. The data used in this study are from the MPS records of individual students. These students’ identities are disguised. But the manner in which they are disguised allows us to track the students by using a unique identifier assigned to each student. Thus, we are able to track students over time, and we are able to associate select pieces of individual information, such as grade, race, gender, and eligibility for free or reduced-price lunch with their standardized test scores in math and reading.

Standardized test results were used as a measure of achievement. MPS uses annual tests (the same test but with two different names) in grades four through ten. This allows annual comparisons to be made rather than having to use only fourth-, eighth-, and tenth-grade tests, the ones mandated by the State Department of Public Instruction. Thus, student identifiers are associated with the Wisconsin Knowledge and Concepts Exam (WKCE) for fourth, eighth, and tenth grades, and the Terra Nova exam for fifth, sixth, seventh, and ninth grades. The exams are given in the fall (usually November) of each year.

We selected two classes to examine in this study: the Class of 2008 and the Class of 2011, each named for the year of scheduled high school graduation. For the class of 2008, fourth-grade WKCE exams were given in 1999, eighth-grade WKCE exams were given in 2003 and tenth grade WKCE exams were given in the fall of 2005. In an attempt to better understand how students actually did at the time they were first scheduled to take the exams, the individual records were created so that the only scores available for analysis were those associated with the first attempt at taking each test. If students are kept behind, they drop from the class of analysis.

The Class of 2011 was constructed in similar fashion. But the data start at second grade because MPS changed its rules and organized testing for second-graders. Thus, their first tests were the Terra Nova, given in 2000. And because they started later, the most recent test scores available are the seventh grade Terra Nova from the fall of 2005.

Results of the exams are expressed in a three-digit number for each subject area. The scores we are using are referred to as “scale” scores.[xx] These scores allow us to follow progress over time because they are “scaled” so that progress or lack thereof can be easily seen. These exams do cover subjects beyond reading and math. But since those are fundamental subjects, they are the ones on which we concentrate.

Each class started with over 7,000 members. The Class of 2011, for example had 7,614 members in 2000, the year they were in second grade. By the time they were in sixth grade the class had diminished to 6,845 students. While the class was smaller by 769 students that is still a large number of students. When we did the analysis, we used information on fewer students because we needed complete records. Thus, when we analyzed the class of 2008 in sixth grade for details on gender, race, and income, we used records from 6,960 students. For the Class of 2011’s sixth grade for a similar analysis, we employed records of 6,407 students.

The vast majority of students in these classes are minority, and the clear majority of these students qualified for free- or reduced-price lunch. Thus, in the sixth grade of the Class of 2008, some 5,031 minority students qualified for free lunch, and another 552 minority students qualified for reduced-price lunch (Table 1). Subsidized students constituted 92% of minority students for whom we had data.  

The concentration of minority students means that the pool of white students is modest in scale to begin with. For this same sixth grade of the Class of 2008, we had data for 930 white students or 13% of the students for whom we had complete records. That is quite close to their overall proportion in the district (14%). When the white student population is subdivided by income, sub-groups, such as those denied subsidized lunches, are not very large. The same is true of Hispanics and even African-Americans. So the most reliable test scores are those for students eligible for free lunch in all three racial groups and whites with incomes too high to be eligible. We use test scores on all sub-populations, but we urge caution drawing firm conclusions on some of the smaller sub-groups, such as those with reduced-price lunches or those denied subsidized lunches.   

METHODOLOGY

The analysis of MPS student achievement is very straightforward. All available test scores were aggregated and divided by the number of appropriate students to create average, standardized, scale-scores for each group of students. Thus, for example, the reading scale scores of all fourth-grade males and females from the Class of 2008 were put together and divided by the number of such males and females to create an average scale score for fourth-grade males and females from that class. Similar figures were created for the fourth-grade from the Class of 2011. The differences in scores between the males and females of each class were compared to see if they are similar.[xxi]

Average test scores were also created for sub-parts of the original gender group. Thus, fourth-grade females were sub-divided into three racial groups—African-American, Hispanic, and white—and then again the three were subdivided into, for example, African-American fourth-grade girls who received a free lunch, those who received a reduced-price lunch, those who applied and were denied a subsidy for lunch, and those who were not eligible for lunch support. This further subdivision was undertaken for all grades, for all three racial groups, for two genders, and for two different classes. All of the averages were compared with comparables to see what patterns exist and to begin to explore what case might be made to address any particular sub-groups.

Averages were chosen because they are an easily understood measure. They can show whether there are small or large differences between groups. They are easily computed. And they give a clear picture over time of how groups of students are doing.

The basic reason for the analysis of the MPS students is to see whether and to what degree conditions in Milwaukee match conditions elsewhere. We expected to see evidence that African-American males, especially low-income males, started behind everyone, including comparable African-American girls, from early years in school, especially in reading. That is what evidence from elsewhere suggests. If true, that then raises the question of what interventions, if any, are needed to address this condition, especially since between 25% and 30% of the MPS student population consists of low-income, African-American males. Discussion of alternative approaches to addressing this issue is what follows the exposition of the scores within MPS.   

ANALYSIS

Gender Differences by Grade in Milwaukee           

Before examining differences by race and income, it is important to first get a picture of how males and females fared on standardized tests, as they move through school. We begin this analysis by choosing one class, the class of 2008 in MPS, in order to be able to follow scores for largely the same individuals from fourth to tenth grade. We will subsequently examine the results for the Class of 2011 that allows us to see the scores for second and third grades to see if the patterns start even earlier in children’s school experience and to see if other patterns are the same for two very different classes.

Table 2 shows the average reading scale scores by gender for all students in the Class of 2008 who took the standardized tests offered at MPS each year from 1998 through 2005. This is basically one cohort, although there are individuals who dropped back into this cohort and members of this cohort who dropped back and took some tests later than their original counter parts. But basically, these are the same or similar individuals. We are largely comparing these class members with themselves, not with a totally different class.

Gender is extremely important to this population’s test scores. In fourth grade, the average female has a test score that is over 6 points higher than the average male. By eighth grade the difference is almost 12 points. By tenth grade, even with a different and lower calibration of scores on the test, the females outscore the males by 20 points, on average. The trend throughout higher grades is a growing disparity between males and females. The pattern is evident by fourth grade. It almost doubles by fifth grade and inches up until 9th grade, when there is another large jump. Females clearly do better than males, on average, on reading. One should note that this MPS class is largely minority and largely low income, skewing the averages in ways that will be examined below. Nevertheless, gender does matter and males do not, on average, achieve at the same levels as females on reading.

The question is whether this pattern prevails in math as well. The research cited above suggests that the genders should be quite similar in math achievement, based on innate ability, but that cultural differences (e.g., greater math emphasis for males) might lead to findings that are the reverse of those in reading.

Table 3 is not as dramatic as the one for reading scores. But females outscore males, on average, in math in all grades between fourth and tenth. The gap is small at first, but it grows in grades six and seven, declines in eighth and ninth and jumps again in tenth. But none of the gaps is very large (six points being the largest difference). Basically, we can say that males do not, on average, outscore females in grades 4-10 on standardized math test scores. The pattern this class exhibits suggests that while males may appear, on average, to be behind on math achievement, the differences are relatively modest. 

Income’s Influence

A second way to look at male and female differences is to see whether there are greater male and female differences among students with different family incomes. The major question is whether MPS faces the same conditions found many other places—that lower incomes are associated with lower levels of achievement and that the pattern of females achieving at higher levels than their male counterparts holds regardless of income. To ascertain this, we will examine students for grades 4-10 for the Class of 2008 for reading and then for math.

Since income levels are not available, we must use a surrogate, eligibility for free or reduced-price lunch. Eligibility for subsidized lunch is based on one’s family’s income being either poverty level or within 175% of federally established poverty levels. Those eligible for free lunch are from the lowest-income families. Those who are eligible for a reduced-price lunch are next-lowest income. Those who have applied for subsidized lunch because they think they may be eligible but are denied are third-lowest income. And those who did not apply are said to be highest income. Some of these students may be eligible for a lunch subsidy, but because of stigma or lack of knowledge, they did not apply. We include them with the non-eligible population because we have no contrary knowledge, and it is not until high school that participation in the subsidized lunch program declines.

Table 4 reveals the average scale reading scores for fourth grade by gender and lunch-eligibility status for the Class of 2008. The first point to note is that on every male/female comparison of seemingly equals, that is the same lunch status, females, on average, outperformed males in all four income categories. The level of income does not change the fact that in MPS females outperform males. Also worth noting is that the scale of the male to female difference is basically the same regardless of income.

The reader should additionally note that as incomes rise from eligible to partially eligible to being close but denied to not being eligible, average scale scores rise. Being eligible for reduced-price lunch or applying but being denied students do better, on average, than low-income students but not nearly as well as those with no support. This pattern is seen in every grade for which data are available (grades 4-10; not illustrated here). Basically, level of income is strongly related to achievement.

Since that income-achievement link is clearly established and since the counts of students in the reduced and denied pools are relatively small, they are excluded from the next table that displays the gender gap for the poorest (free lunch) and those best off (no support) across grades 4 through 10 (Table 5). The first point to note is the size of the reading gap, starting in fourth grade. The gap almost doubles in fifth grade and remains in that vicinity until eighth grade. Ninth grade had a decline in the gap between genders among the poorest students but a huge gain among those with no support. Ninth grade is an anomaly and likely affected by students coming to MPS for high school from private K-8 schools. In tenth grade both income levels have substantial gaps. The basic trend is a gain in the size of the average reading gap within each income category as the students move to higher grades regardless of income level.

More important to note is the scale of the difference between those eligible for free lunch, the lowest-income students, and those not eligible for lunch support. The differentials are very large, be it among males or females (Table 6). For example, at fourth grade, females with no lunch support scored an average of 33 points (658-625) above females with free lunches. For males the difference was the same, 33 points (652-619). The difference between income levels within the same genders is large but relatively similar across most grades. But in high school the gap enlarges for females in ninth grade and males in tenth grade. The initial pattern grows over time: on average low-income males fall further and further behind both middle-income males and middle-income females.

Low-income males were 39 points behind the non-subsidized females, on average, in fourth grade reading. By fifth grade the gap was 46 points. By sixth grade it was 51 points. And by tenth grade low-income males were 70 points, on average, behind middle income females in the District. Those are very substantial differences. While it is very difficult to be precise, males could be interpreted as being between one and two years or more behind middle-income females.[xxii] Gender matters, but even more pronounced is the effect of income.

Math score differences should not be as pronounced, given the relatively small differences seen above between genders across the seven grades. Table 7 shows the scores for the fourth grade of the class of 2008. The gender pattern is the same for three groups, those eligible for free or reduced-price lunch, and those not seeking support: females outperform males. The difference among those denied support is nil. Regardless, there appear to be extremely modest average differences between genders within income groups.

When we examine just the lowest-income and middle-income students by gender, we see a very mixed series of numbers (Table 8). There are occasionally larger gaps between those on free lunch, but for the most part the differences are modest. Basically, gender differences do still exist on math scores, and in most instances the differences in averages are small within the different income groups.

Table 9 displays the differences in average math scale scores between low-income and middle-income females in the first column.

 In the second column is the difference between low-income males and middle-income males in MPS by grade. The third column shows the average score differential between the low-income males and the usually higher-scoring, middle-income females in the same grade. The main message is that within each grade, the differences between students with no support and with free lunches are not quite as large as on reading. But they are still substantial. In fourth grade, the two pools of females differ, on average, by 33 points and the males differ by 27 points. By tenth grade the average gap between higher and lower incomes is 37 points among females and 42 points among males. Those are extremely large differences and translate into at least one and perhaps two or more years of achievement, on average.

The third column in Table 9 is included to point out just how far low-income males are behind middle-income females, on average. The numbers are almost always larger than the ones that show the gap between low- and middle-income males. The gaps are usually not as large as on reading, but they are very large.

.

Reading Scores by Gender and Race

Given the literature that pinpoints race as a critical factor associated with differences in levels of student achievement, we need to examine scores by subject, by grade, and by race to learn the situation in Milwaukee. The data to aid this exploration are available. They are examined first with all incomes combined and then differentiated by lunch status. Obviously, in Milwaukee with 74% of students eligible for free or reduced-price lunch in 2004, the picture given by a view of all incomes combined will be strongly influenced by the dominant low-income population.

Table 10 needs to be examined across both gender and race. Looking at gender for fourth grade for all three groups sets the general stage. Males, on average, are a bit behind females among African-Americans, Hispanics, and whites. That pattern does not change as the class aged and moved up through the grades. The one difference is that males are a bit further behind their female counterparts as they progressed through school. Thus, African-American males in ninth grade are 16 points, on average, behind African-American females while they were only 8 points behind in fourth grade. White males in ninth grade are, on average, 16 points behind white females while the males were only 4 points behind in fourth. Among Hispanics, the four-point gap that existed in fourth grade averages grew to 12 points by ninth grade. The pattern is clear: males, on average, do not do as well in reading as females, regardless of race. But the gap is most often greatest among African-American students.

A second important way to read this information is to compare racial differences to learn of differences within the same gender across racial groups. As expected, at every grade, white females had substantially higher average scores than Hispanic females. And both white and Hispanic females scored higher, on average, than African-American females. The same pattern holds for males.

These patterns are spelled out in Table 11. The differences between average African-American females and average Hispanic and white females are detailed. Most often the gap with Hispanics is in the 6- to 10-point range. But between African-American and white females in the same grades, the differences are often 26 to 30 points and jump to 52 points in 10th grade. Similar patterns are visible among the males, only the Black/white gap is 56 points in 10th grade.[xxiii] Thus, it is clear that African-American males have been achieving at lower levels, on average, in reading than their cohorts of other races within MPS, be they males or females. The African-American male is, on average, behind by fourth grade and continues to be even further behind as the class moves through subsequent grades.

The next logical question to explore is whether males are behind females in math, as they are in reading. Table 12 shows the distribution of average scores by grade, gender, and race for the three largest student groups in MPS. A glance at the first line, fourth grade, reveals that math is different from reading. Only African-American males have average scores lower than African-American females. Whites and Hispanic males and females have the same scores. Again, there is evidence that African-American males are behind others early in their academic careers.

Gender differences on average scores are quite small, ranging from 0 to 4 points between males and females by grade among whites and Hispanics. For these two groups, there is not, on average, a difference in math achievement across genders. But for African-Americans there are larger differences. The differences are not consistent but grow from 4 points at fourth grade to 10 at tenth grade, hitting 8 points in sixth and seventh grades before dropping to 3 points in the mix of ninth grade. African-American male averages always are below African-American females. The scale of difference in math does not match that in reading, but it does exist and must be noted. More critical are the racial and income differences that exist.

What also should be noted is that the average scores of both males and females among African-Americans are quite far below those of Hispanics and whites. This pattern is likely largely related to income, the subject of the next section.

Before going to income discussions, the scale of differences should be examined by grade and race (Table 13). African-American males in fourth-grade math are, on average, 12 points behind Hispanic males. By eighth grade the difference in averages is 21 points, and by tenth grade the difference is 27 points on an exam scale that understates the difference relative to the 4th-9th grade scores. African-American males not only are behind in math by fourth grade; they drop further behind as they age. Hispanic males and females are about half as far behind whites as African-Americans, on average.

The gap between the fourth-grade, African-American males and the white males is, on average, 28 points. By tenth grade, the difference is 53 points. That is likely more than two years of average progress and indicates a severe difference in achievement in math. This is not new news, but its repetition points out how difficult the challenge of raising math test scores, when the gap is so large in the early years of elementary school.

--->Next Page (1...2)