CHAMPAIGN, Ill. - Policymakers may want to rethink how they determine when children with limited English skills are fluent enough to learn in English-only classrooms, says a new study by an education professor at the University of Illinois.
The more than 5 million children in America's schools who aren't native speakers of English are the fastest growing segment of the school-age population, and, according to some projections, may compose more than 40 percent of elementary and secondary students in the U.S. by 2030. Determining when non-native speakers have enough English mastery to learn in mainstream classrooms - a transition termed reclassification - is crucial to ensuring students' academic success as well as schools' abilities to achieve the learning benchmarks mandated by the No Child Left Behind Act.
In a new study, Joseph P. Robinson, a professor of educational psychology in the College of Education at the U. of I., examined policy-based thresholds for reclassification used by a large urban California school district, where English language learners account for about a third of the student population. By using a more rigorous statistical method than in previous studies of reclassification, Robinson found that the district's existing policy threshold was reclassifying some students too early.
While policies on reclassifying ELLs vary widely across school districts and between states, California's education code suggests that school districts use multiple criteria to assess students' readiness for reclassification as fluent English proficient.
In the California school district that Robinson studied, ELLs must attain minimum scores on five tests, including a score of at least 300 on the state English language arts test, in order for a faculty committee to consider them for reclassification. Falling short on any one of those five scores - even by a single point - means that a student has virtually no chance of being reclassified at that time. Yet a peer whose scores are right on the threshold has about an 80 percent chance of reclassification, Robinson found.
"What previous studies have done is statistically adjust for students' prior academic achievement and then see what the difference is between those students who are reclassified and those who aren't," Robinson said. "The problem with that type of analysis is selection bias: We don't know why, if one kid scores the same as another, one of them is reclassified and another one isn't.
"The advantage to the method I'm using is that, in terms of their mastery of English, there's really no difference between a kid who scores 299 and a kid who scores 300, except one of them likely will get reclassified and one of them won't. Looking at what's going on right at that critical point allows us to assess these policies while also eliminating selection bias."
If the assessment threshold is set at an appropriate level, students' academics should show no effects - positive or negative - after the transition, Robinson said.
"From a methods perspective, the traditional regression analysis - a common method used in prior studies - leads you to believe there's a significant positive effect of reclassification in the year that the student is reclassified, and even the year after, in elementary school, middle school and high school," Robinson said. "If policymakers are basing their decisions on that type of analysis, they're going to reach the conclusion that perhaps the threshold should be moved downward so more students are reclassified sooner. That's the conclusion that most of the previous studies on reclassification came to."
However, Robinson noted that regression analyses are often biased by factors that may have influenced the final decision to reclassify a student, such as teachers' observations and parental opinions, which are seldom officially recorded and therefore cannot be taken into account statistically.
Using a form of regression discontinuity design, an analytical technique used by social scientists to develop plausible estimates of treatment effects for individuals near a cutoff score and to determine causality, Robinson found significant negative effects of reclassification on ninth- and 10th-graders near the policy threshold the year after they were reclassified.
"What this more rigorous type of analysis is saying is, 'No, the current threshold is resulting in worse outcomes for the reclassified students.' An option for policymakers is to raise the threshold higher so that students are reclassified later."
Or, educators may want to re-examine the base curriculum to ensure that it is preparing ELL students for advanced material, or look at teacher preparation for instructing students with limited English to see if it is affecting students' outcomes, Robinson said. "This study can say whether there is strong evidence to be concerned about where your policy threshold is set. However, it doesn't tell you about the specific mechanisms for how you could remedy that."
Reclassification as fluent English proficient has far-reaching consequences for students, including reductions in or elimination of English language development services and immersion in a new environment with different peer groups and teachers, who may have little preparation for instructing students with limited English, Robinson said.
"When students are reclassified, much changes for them all at once, and you can't disentangle all of those factors with this type of study," Robinson said.
However, the study is "a big step forward" in that it helps isolate the effects of policymakers' guidelines from the highly specific decisions made by teachers and parents.
Robinson also found that the district's reclassification criteria had no effects on other outcomes, such as student attendance, enrollment opportunities for advanced courses or students' successful completion of any academic subject.
Robinson cautioned that the results are specific to the school district studied and may not be generalizable to other schools, since districts' policies are idiosyncratic.
The study appears in the September issue of the journal Educational Evaluation and Policy Analysis.