What Doesn't Meet the Eye

Does Socioeconomic Status Predict Achievement Disparities?

The analysis in this section is designed to answer two related questions. One question is whether SES helps to predict racial and ethnic differences in achievement. Many studies have addressed this question, and the answer is virtually always yes.¹⁵ The other question is whether the magnitude of the achievement gap is different for different levels of SES. For both questions, the answer here is yes. Disparities in SES predict substantial portions of the disparity for each measure of achievement, but not all of it.¹⁶ In addition, the residual "unexplained" disparity, holding SES constant, is greater among students from high-SES households.

For this analysis, the grade-point average from the most recent term is measured on a four-point scale; the other two achievement variables from Table 1 are measured now in standard deviation units.¹⁷ Also, it is worth noting that the SES variables here relate conceptually to home intellectual resources (such as books in the home, computers in the home, and parents' education) and number of parents per child (number of siblings and number of parents). The data for this study lack financial-status measures, such as wealth, income, or free and reduced-price lunch status.¹⁸

The analysis here uses four standardized SES categories: lowest SES, lower-middle SES, upper-middle SES, and highest SES.¹⁹ Table 3a shows what percentage of each race/ethnic group is in each of them. Only 2 percent of blacks have SES characteristics in the highest SES category, while only 3 percent of whites have characteristics in the lowest category. Approximately 79 percent of blacks, 78 percent of Hispanics, 56 percent of mixed-race students, 46 percent of Asians, and only 28 percent of whites are in the lowest and lower-middle categories combined. A look back at Table 2 shows the types of SES disparities for particular variables that together account for the disparities in Table 3a.

Table 3b uses SES "profiles" constructed from the SES categories mentioned above. For a given SES category, say "lower-middle SES," the SES profile comprises the list of mean SES characteristics across all race/ethnic groups combined. Thus, each profile is identical for all race/ethnic groups in a given SES category.

Table 3a Percentage Distribution of Each Race/Ethnic Group Across Four SES Categories Percentages SES Category Black White Hispanic Asian Mixed Race Total Lowest SES 24 3 19 7 12 10 Lower-Middle SES 55 25 59 39 44 40 Upper-Middle SES 19 57 19 41 37 40 Highest SES 2 16 3 12 8 10 Column Total 100 100 100 100 100 100

Table 3b Simulations by SES Profile and Race/Ethnicity for Three Achievement Measures* Panel A: Simulated Mean GPA (4-point scale) SES Profile Black White Hispanic Asian Mixed Race Lowest SES 2.38 2.52 2.61 2.66 2.30 Lower-middle SES 2.65 2.91 2.88 3.07 2.73 Upper-middle SES 2.88 3.36 3.13 3.36 3.17 Highest SES 3.18 3.68 3.34 3.67 3.49 Panel B: Simulated Amount That the Student "Completely" Understands of Teachers' Lessons (Standard Deviation Units) SES Profile Black White Hispanic Asian Mixed Race Lowest SES -0.38 -0.54 -0.44 -0.58 -0.59 Lower-middle SES -0.23 -0.22 -0.21 -0.26 -0.26 Upper-middle SES 0.00 0.20 0.01 0.06 0.22 Highest SES 0.04 0.35 0.11 0.35 0.31 Panel C: Simulated Amount That the Student Understands "Very Well" of Material Read for School (Standard Deviation Units) SES Profile Black White Hispanic Asian Mixed Race Lowest SES -0.56 -0.59 -0.65 -0.64 -0.57 Lower-middle SES -0.36 -0.15 -0.39 -0.29 -0.31 Upper-middle SES -0.07 0.25 -0.06 0.17 0.17 Highest SES 0.06 0.44 0.17 0.41 0.36 * Simulations are for fixed SES profiles, where achievement predictions use regression coefficients estimated separately by race/ethnicity. See text and footnotes for more detail.

The "prototypical student" defined by a given SES profile has a different predicted achievement level, depending on race/ethnicity. This situation is true for each of the three achievement variables (grade-point average, comprehension of lessons, and understanding of readings). The lowest SES level shows the least race/ethnic achievement disparity.²⁰ For this profile, the predicted black-white gap in grade-point average (GPA) is only 0.14 GPA points; the predicted grade-point average for Hispanics is actually 0.09 points higher than for whites (Panel A of Table 3b). Similarly, in Panel B and Panel C of Table 3b, the other two achievement measures do not show any clear tendency for whites to rank higher than other groups. Generally, these findings show only small race/ethnic achievement gaps in MSAN districts among students with the lowest SES profile.

However, at the highest SES level, the disparity among groups is much greater. Whites rank highest and blacks lowest, with sizable gaps between them. The predicted grade-point average gap at the highest SES level is one-fifth of a GPA point between whites and mixed-race students, one-third of a point between whites and Hispanics, and a full half-point between whites and blacks. The rank order of predicted achievement among groups is the same for the two skill measures in Panel B and Panel C. Note that the predictions for whites and Asians are essentially equal across all three measures.

High-SES students achieve at higher levels than middle-SES and low-SES students among all racial and ethnic groups. However, findings here indicate that the degree to which SES pays off differs among groups. For all three measures, the difference in achievement between high- and low-SES students is smallest for blacks and Hispanics. The reasons are not entirely clear and will be the subject of ongoing research by this author and others. The differences may simply be artifacts of the (in)accuracy with which students answered the survey. More likely, these differences may reflect race/ethnic differences in home, peer, and classroom processes among high-SES students. In any case, it appears from this analysis that SES differences (and the differential life experiences that they represent) account for some but not all racial and ethnic differences in student-reported grade-point average, understanding of teachers' lessons, and comprehension of materials read for school. Further, the unexplained racial differences are greatest at the highest SES levels.

Previous | Table of Contents | Next