School Finance 104

My post yesterday discussed the various ways in which additional funding might be allocated for specific cost factors. In this post, I want to discuss issues related to weights for student poverty and English Learners, the two student cost factors explicitly included in Brown’s LCFF. 

 

The most common categories of student need that states include in their school funding formulas are special education, at-risk students (generally meaning low-income but may also mean any needing remedial education) and English learners (ELs). In many states, the amount allocated for these higher-cost students is determined through pupil weights, set as a percentage of the base allocation (see last post). If a student falls into multiple need categories, the weights might be added together, or not (Brown’s formula takes the latter approach). Some states (e.g., Oregon, Florida) cap the total weight that can be assigned to an individual student. 

 

The LCFF calls for a weight of 0.35 (35% more) for low-income students or English Learners, and the weight increases if the concentration of weighted pupils in a district exceeds 50%. In general, weights for these categories reflect the research on the costs of education, which consistently finds that poor and EL students require additional resources in order to achieve at the same levels as students who do not face these disadvantages. There is also relative consensus on at least a range for the magnitude of appropriate weights. Research is less clear whether/how these costs change as the concentration of disadvantaged students increases.

 

Imazeki (2007) synthesizes the estimates of the marginal cost for poverty and English learners from 16 costing-out studies (i.e., studies that attempt to estimate the costs of achieving an ‘adequate’ level of education). In add-on pupil weight terms, the estimates for poverty range from 0.30 to 1.22. The estimates for English Learners range from 0.24 to 1.01. In addition, much of the research that has been in done in various states on the costs of services for English Learners finds marginal costs in the range of 20 to 30 percent. A large determinant of costs is the instructional approach used (e.g., separate ESL instruction versus immersion). The Getting Down to Facts cost studies done specifically for California (Chambers et al. 2007, Sonstelie 2007, and Imazeki 2007) all establish pupil weights for poverty of at least 30%. 

 

Among the roughly 30 states that use a weight or adjustment for low-income students in their school funding formula, those weights range from 5% (Mississippi) to 50% (Maryland), with the majority falling in the mid-20’s (as of 2007; see Verstegen, 2011). A few more states have weights or adjustments for English Learners but the size and form of those adjustments are much more varied (e.g., several states establish a dollar amount per EL, rather than a percentage weight, while others establish a total state appropriation for ELs that is distributed proportionally to districts, and some states use a combination of approaches).

 

The previously-mentioned studies typically report a single weight for all poor or EL students. An alternative approach is for the weight to vary with the concentration of disadvantaged students in a school or district. This approach is based on evidence that peer effects matter, i.e., that poor students in schools with low concentrations of poverty perform at higher levels than similar students in schools with high concentrations of poverty. The research is less clear on the threshold where concentration begins to matter (i.e., the point above which we would want to give additional funding), nor the magnitude of the effect so it is unclear how much the weight should increase as concentration increases. There is some evidence that concentration may be less important for English Learners (e.g., Imazeki and Reschovsky (2004) find costs fall as EL concentration increases in Texas); this could be due to economies of scale in providing specialized instruction but the relationship between costs and EL concentration has not been well-studied. 

 

Very few states adjust weights specifically for concentration of poverty. A handful restrict EL funding to districts above a particular threshold; for example, Connecticut only allocates the EL supplement to districts with more than 20 or more EL students. 

 

Having separate weights for poverty and for English Learners raises questions about how to handle students who fall into both categories. Gandara and Rumberger (2007) conclude that although English Learners may require different resources than low-income students, it is not clear that they need more. This is consistent with the conclusions of Rose, et al (2008) that while low-income students and EL students each require additional resources (relative to students who are neither), students who are both do not require additional resources on top of that. Their data also suggest that it is poverty more than language ability that drives the additional costs. This suggests that in a formula with weights for both poverty and ELs, students who are both should only receive the weight for poverty. However, among the other states that provide funding for both low-income students and English learners, all of them ‘double-count’ students who fall into both categories, although some states do cap the total possible weight (from all categories) assigned to a given student.

 

Some policymakers have raised concerns that by allocating additional funds to districts with English learners, districts will have a disincentive to re-classify students who gain English proficiency (as this will lead to a loss of revenue for the district). This argument is presumably based on a similar argument often heard about classifying learning disabilities which generate special education revenues. However, there is no research evidence on whether or not this is a problem for English Learners. It should be noted that California already allocates at least some funding based on EL counts (e.g., ELAP, EIA) but it is not clear whether this has had any impact on re-classification rates. Still, Brown’s LCFF restricts funding for ELs after five years.

 

Perhaps the most-repeated argument for keeping categorical restrictions on funds for low-income students and English Learners is that without those requirements, districts will stop providing the extra attention these students need. On the one hand, such concerns are understandable, given the persistence of achievement gaps, and given the top-down regulatory culture that has permeated California school finance for thirty years. Back in the 70’s and 80’s, the easiest way to make sure districts were helping their most disadvantaged students was for Sacramento to mandate that they do so and withhold funding if they didn’t. But today, things are very different. We have an accountability system that explicitly recognizes different student groups, and we also have a political environment that is far more invested in the progress of all students (although we still have a long way to go). The progress that has been made in the last decade to narrow achievement gaps (however small and slow) has not been because of categorical restrictions, but because the assessment and accountability system makes those gaps so much more transparent than they used to be. Replacing categoricals with unrestricted weights, coupled with strong accountability standards, is likely to free districts to find more effective ways to help their poor and EL students, not abandon them. 

 

Overall, Brown’s proposed weight for low-income and English Learner students seems reasonable when compared to the research and experience in other states. 35%  is a bit on the high side for a marginal weight for each of those categories individually; however, the fact that poor and EL students are not double-counted, and that there is a large overlap in those populations, mean that the resulting allocations could end up looking similar to what we see in other states, where the weights are lower but low-income and EL students are funded separately.