June 26, 2008

The Role of Schools in the English Language Learner Achievement Gap

III. The Mathematics Pass Rate of English Language Learners

In the wake of the NCLB legislation, several recent analyses have examined the measured school achievement of English language learners (ELLs) at both the national and state level (Batalova, Fix and Murray, 2007; Fry, 2007). In both reading and mathematics, a majority of ELL students who took assessment tests scored below proficiency standards. Furthermore, ELL test-takers were far behind the other major racial/ethnic groups in measured achievement in elementary school. The gaps with other major racial/ethnic groups widen from elementary grades to eighth grade.

For the five states with large ELL student populations studied in this report, Figure 1 illustrates the difference in the share of white test-takers who scored at or above each state’s proficient level in mathematics and the share of ELL test-taking students who scored at or above the proficient level (or the difference in the “proficiency rates” on the mathematics assessment). For example, according to the 2004–05 NLSLSASD, 49% of Arizona grade 3 ELL test-takers met or exceeded the Arizona math standard on Arizona’s Instrument to Measure Standards (AIMS).1 Nearly 84% of grade 3 Arizona white test-takers met or exceeded the standard. Thus, Arizona ELL test-takers trailed their white counterparts by 35 points, or the difference in white and ELL grade 3 math proficiency rates was 35 points in Arizona.2 In all five states, the measured mathematics achievement difference increases from the early elementary grade (typically grade 3) to the middle school grade (typically grade 8).

In the 2004–05 NLSLSASD, the ELL test-takers often trail black test-takers in measured math proficiency (Figure 2). In grade 3 math in California and Texas, ELL test-takers were more likely than their black counterparts to meet or exceed the state standard, but otherwise ELL test-takers trailed their black peers. Fry (2007) reported large differences between standardized test scores for black and ELL students, based on an analysis of data from the 2005 National Assessment of Educational Progress (NAEP).

  1. numoffset=”5″ The NLSLSASD, being a school-level data set, records the percentage of test-takers in each public school who meet or exceed state-designated assessment cut points. The state aggregate “proficiency rate” for group i is simply the weighted average of the public schools’ proficiency rates for group i, where the weight is the number of group i’s test-takers in a public school. That is, school proficiency rates with larger numbers of group i test-takers receive more weight in state aggregate “proficiency rates.”
  2. It should be noted that when comparing ELL test-takers to the major racial/ethnic groups of test-takers, one is not comparing mutually exclusive groups. That is, ELL status is not a racial/ethnic category. An ELL test-taker’s performance is included in both the ELL group and one of the major racial/ethnic groups. ELL is a public school-designated status, and school administrative data sources do not reveal the race or ethnicity of ELL test-takers. So, fundamentally, we do not know the exact overlap between the ELL test-takers and racial/ethnic groups of test-takers. However, Census Bureau data suggest that the overlap between public school students with limited English-speaking abilities and white public school students and black public school students is quite minimal. Nationally, only about 1 percent of white public school students have limited English-speaking abilities. A similar percentage applies to black public school students. This suggests that nationally most white and black test-takers are not also ELL test-takers. Granted, in the five large ELL states examined in this report, there is probably more overlap between white and black test-takers and ELL test-takers than is the case nationally, but it is still likely not to be extensive.