University of Arkansas Office for Education Policy

English Language Learners in Arkansas

In The View from the OEP on June 18, 2019 at 3:16 pm

English language learners (ELs) are a growing student population throughout Arkansas.  With the passage of ESSA, the progress of EL students towards proficiency in English has been added to the metrics by which schools and districts are held accountable. With this academic context in mind, we at the Office for Education Policy thought it would be useful to provide a descriptive analysis of the EL population in Arkansas. These analyses, beginning with an overview of the recent enrollment trends in the state, may help policymakers and school leaders better understand this important group of students.

Recent data from the National Center for Education Statistics (NCES) show that ELs comprise about 9.5% of K-12 students in the United States. Some states have a larger share of EL students than others. For example, in fall 2016, only about 1% of West Virginia’s student population were ELs while nearly 16% of students in Nevada were ELs.

We really like this map from USDOE that shows the percentage of ELs by school district across the country.

EL Map

Additionally, some states are experiencing faster growth in the proportion of EL students than others. For example, the states with the highest growth of EL student population from 2000 to 2014 include Arkansas.

The change in Arkansas’ EL population by district from 2009-10 to 2014-15 is presented below.  Most districts had few or no ELs, as represented by black or gray shading, but many districts are colored dark purple, indicating an increase in EL enrollment of at least 50% over the years examined.

El Change legend

Change in EL

 

With such large changes in the EL population by 2014-15, we wanted to bring in more current data to examine the trends here in Arkansas.  Examining data from the Arkansas Department of Education, we describe overall EL enrollment in the state, the proportion of EL students to general population, and the proportion of EL students by academic region for the past 15 years.

Table 1, below, presents the number of students enrolled in English language learning programs over the last 15 academic years. Overall, the total number of EL students enrolled in Arkansas schools has doubled from just under 20,000 students in the 2005 academic year to just under 40,000 in the 2019 academic year. All regions have seen increases in the number of EL students from 2005 to 2019. In total numbers, the Northwest academic region welcomed over 23,000 EL students in AY 2018-19 – the most by far.

Table 1. EL Enrollment in Arkansas, 2004-05 through 2018-19.

ELinAR

Changes in the proportion of EL students in Arkansas from 2004-05 through 2018-19 are illustrated in Figure 1, below. The figure shows the EL percentage of the general student population in Arkansas, by academic region, and in the United States overall. The available data for the national EL rate, represented by the blue dotted line, reflect that the overall share of the EL K-12 population is consistently between nine and ten percent of the general school population.

Figure 1. Percentage of General Population Identified as EL, 2004-05 through 2018-19.

ELProportion

Several key points should be noted from Figure 1.

  1. Although Arkansas enrolls a smaller percentage of EL students than the national average, the share of Arkansas public school students identified as EL is increasing, from 4% in 2004-05 to 8% in 2018-19.
  2. While Northwest Arkansas consistently enrolls the highest percentage of EL students in the general student population, both the Southwest region and the Central region have seen large relative increases in the percentage of EL students enrolled.
  3. A recent decline in the percentage of EL students is also evident in Figure 1. Starting with academic year 2016-17, there is a downturn in the percentage of EL students in Arkansas overall, and a sharp decline in the Northwest region in particular. This decline coincides with a change in the assessments used to reclassify EL students as proficient in English. Beginning in academic year 2016-17, Arkansas replaced the ELDA with ELPA21, and although not evident in all regions, the decline in identified students overall is likely due to a greater percentage of EL students being identified as English Proficient through the new assessment.

In addition to looking at the proportion of ELs enrolled in the general population, we examined which region has the highest proportion of EL students overall and how that has changed over time. Figure 2 illustrates the shift in the overall share of EL students between the academic regions over the last 15 years. While all regions demonstrate growth, Figure 2 illustrates that while Northwest Arkansas consistently has the highest number of EL students enrolled, the region’s proportion of the EL population in the state is decreasing, due, in part, to an increase in the numbers in the Central region.

Figure 2. Share of EL Enrollment by Academic Region, 2004-05 through 2018-19.

ELALLRegionIn 2004-05, over three-fourths of the state’s EL students were enrolled in Northwest Arkansas schools. This has dropped to 61%. In contrast, both the Central and Northeast regions have experienced growth in the share of the overall proportion of EL students. The Central region’s share, for example, has increased from 11% to 2005 to over 21% in 2018-19.

We finally illustrate this major shift in student composition in the state by looking at the extremes – districts with no EL student enrollment and districts with at least 10% of their student population identified as ELs. We illustrate this in Figure 3. While the number of EL students in Arkansas has increased, so has the number of individual districts that enroll these students. Look at the Northeast region, for example. In 2004/05, 56 of the 75 districts in the Northeast region did not enroll a single EL student. None of the remaining districts in the Northeast region had EL enrollments that comprised 10% of the overall student population. By 2019, however, only 17 districts in the Northeast have no EL enrollment and two districts had EL enrollments that comprised over 10% of their student population. In 2019, every academic region of Arkansas has at least two school districts where at least 10% of their student population are English learners.

Figure 3. Share of EL Enrollment by Academic Region, 2004-05 through 2018-19.

ELExtremes

legend


The English Learner student population in Arkansas has more than doubled over the last decade and a half. While the majority of EL students attend schools in Northwest Arkansas, the Central and Southwest regions have higher rates of growth. Additionally, schools in the Northeast academic region account for whittling some of the Northwest region’s share of EL students. The share of EL students in the Northeast has grown by over 2.5 percentage points from 2018 to 2019.

While the number of EL students has grown, so has the number of districts in which they are enrolled.  In 2004-05, 155 districts throughout the state enrolled no EL students, but in 2018-19 there were only 44 such districts.  In addition, over 30 districts have at least 10% of their students identified as EL, compared to fewer than half that 15 years ago.

The differential growth patterns in the EL population in Arkansas may be problematic for districts that have had few or no EL students enrolled. Districts that have long welcomed EL students may have greater institutional knowledge to identify and support their EL students as they learn and become proficient in English. Districts that have experienced high EL growth, or have recently received their first EL students, may struggle with how best to meet the learning needs of these students. This can be an opportunity for the veteran EL districts to be pedagogical leaders, sharing their expertise with other districts.

It is important to ensure that EL students who attend schools in all regions of Arkansas are afforded the opportunities to succeed in learning English as well as core academic content. In addition to looking descriptively at the educational contexts EL students find themselves in, we will examine how EL students are doing academically in future installments of this OEP blog – watch this space!

 

 

Comparing Traditional School and Open-Enrollment Charter School Expenditures

In The View from the OEP on August 8, 2019 at 7:43 am

Back in May we dug into the brand-new school-level spending data in two blog posts. In our first post we showed that Arkansas’ schools spend most of their money on people (76%) and instruction (56%). We also showed that schools with higher poverty levels, as measured by Free and Reduced-Price Lunch (FRL) percentage, tend to spend more.

Our second post broke the spending data out by school level (i.e., elementary, middle, high). We found that Arkansas spends far more per pupil on high school than it does on elementary and middle school (i.e., >$1,000 more per pupil). Somewhat surprising to us, we also found that Arkansas spends the least on middle schools.

In this post, we compare school-level spending between traditional schools and open-enrollment charter schools. For the remainder of the post, we will refer to open-enrollment charter schools simply as charters or charter schools. Arkansas also has district conversion charter schools, which are traditional schools that have been provided with additional flexibility, but they are funded in the same way as traditional schools. Conversion charter schools are treated as traditional schools for our analysis.

Both traditional schools and charter schools receive funding through the state funding formula. However, local property taxes cover a portion of traditional schools’ per pupil funding amount as determined by the formula, and traditional districts have the ability to raise additional dollars if voters approve higher property tax rates. State-level taxes fill in the gap between the formula funding amount and the amount traditional districts raise through property taxes, increasing funding equity between wealthy and poor districts.

Charter schools, on the other hand, do not have geographic attendance zones, and thus cannot raise local property tax revenue. Charters’ per pupil formula funding amount is fully covered through state-level taxes.

The inability of charter schools to access local property taxes can create big differences in expenditures between traditional schools and charter schools. Facilities funding is often a big driver of these differences. A large proportion of traditional districts’ funding for facilities comes from local property taxes, and state and philanthropic funding for charter facilities generally falls short of the amount traditional districts spend on building projects. This post will not dig into the particulars of Arkansas school facilities funding, but will instead focus overall differences and the expenditure categories we presented in our earlier posts. In the coming weeks, will likely follow up with a post devoted to charter facilities funding.

Arkansas currently has 25 open-enrollment charter school districts comprised of 49 individual schools. Charters make up only about 5 percent of schools in the state, educating around 3 percent of students. Figure 1 below shows a scatter plot of school-level spending versus the percentage of students eligible for the Federal Free and Reduced-Price Lunch (FRL) program. Charter schools are represented by the orange circles. Charter schools are scattered throughout the plot, and do not deviate significantly from the overall spending vs. FRL trend we showed in our first post.

Figure 1: Scatter Plot of 2017-18 School-Level Spending Per Pupil vs. Percent FRL

Screen Shot 2019-08-07 at 11.43.21 AM

While the scatter plot above shows that charter schools closely resemble traditional schools in terms of expenditures and percent FRL, there are some differences in spending levels. Figure 2 illustrates the distribution of school-level spending for both traditional and charter schools. In the 2017-18 school year, the median charter school spent $521 less per pupil than the median traditional school. The difference in the median school’s per pupil spending is largely driven by the 6 charter schools (12 percent of charters) that spent less than $7,000 per pupil, otherwise the distribution of spending per pupil looks similar between traditional and charter schools.

Figure 2: 2017-18 Total Spending Per Pupil

Screen Shot 2019-08-07 at 11.41.18 AM

While overall expenditures look quite similar, there are interesting differences in how charter vs. traditional schools spend their dollars. As we did in our earlier posts on school-level spending, we can also break spending out by activity/purpose. Figure 3 shows the distribution of per pupil spending for instruction, which includes things like teachers’ salaries, books, etc. In the 2017-18 school year, the median charter school spent $4,364 on instruction while the median traditional school spent $5,520, a difference of $1,156 or about 20 percent compared to the median traditional school. Since the majority of instructional expenditures are for teacher salaries which are linked to experience, it’s likely that charter schools’ lower average teacher experience is a big driver of this expenditure difference.

Figure 3: 2017-18 Per Pupil Spending for Instruction

Screen Shot 2019-08-07 at 11.41.43 AM

Charter schools spent less on instruction, but more on administration. Figure 4 provides the distribution of per pupil spending on administration. The median traditional school spent $760 per pupil on administration, including both school and central administration costs, while the median charter spent $1,241, a difference of $481 or about 60 percent compared to the median traditional school. The difference in administration expenditures is not very surprising given the fact that several charter schools are still starting up and the median charter school is about 40 percent smaller than the median traditional school. New charter schools generally start small and grow over time, but school leadership needs to be in place from the beginning. Because administration costs are spread over a smaller student base early on, administration costs are often higher per pupil in a school’s startup phase.

Figure 4: 2017-18 Per Pupil Spending for Administration

Screen Shot 2019-08-07 at 11.41.57 AM

Charter schools spent more on student support services like physical and mental health services, guidance counselors, etc. and less on instructional staff support like teacher professional development. Figure 5 shows the distribution of expenditures on student support services. The median traditional school spent $483 per pupil and the median charter spent $398, a difference of $85 or 20 percent compared to the median traditional school.

Figure 6 illustrates the distribution of expenditures for instructional staff support. The median traditional school spent $716 per pupil on instructional staff support while the median charter spent $438, a difference of $278 or 40 compared to the median traditional school.

Figure 5: 2017-18 Per Pupil Spending for Student Support Services

Screen Shot 2019-08-07 at 11.45.16 AM

Figure 6: 2017-18 Per Pupil Spending for Instructional Staff Support

Screen Shot 2019-08-07 at 11.42.25 AM

Overall traditional school and charter school expenditures per pupil are quite similar; however, there are interesting difference in how they spend those dollars. In the 2017-18 school year, charters spent somewhat less on instruction, owing to their younger, less experienced workforce, but they spend more on administration. Charters also spent more on student support services but less on instructional staff support.

While the state funding formula largely determines the amount of funding schools receive, districts and schools have a lot of discretion around how those dollars are spent. The distribution charts above show that schools spend their dollars in a variety of ways, and the differences across traditional schools are generally bigger than the differences between traditional schools and charter schools. There is no “right” way to spend education dollars. The context and the student population matter a lot. That’s why we are big fans of local decision making as long as it is strategic and student focused.

Why are ACT Aspire scores flat?

In The View from the OEP on July 31, 2019 at 11:42 am

We have been wondering why Arkansas’ most recent ACT Aspire scores are essentially the same as they have been for the past two years.   As seen in Figure 1, the preliminary scores from the spring 2019 ACT Aspire show that statewide, student performance is flat, with fewer than half of Arkansas’s students meeting grade-level expectations in math, science, and reading.

The lack of improvement in these scores is confusing because funding for Arkansas public schools continues to increase (currently $6,899 per student annually),  class sizes are small (an average of 15 students per class), and teachers are being provided professional development, particularly in how to teach reading.  We don’t have any other states to check in with and see how they are doing on the ACT Aspire since we are the only state administering it right now.

Figure 1. Percentage of Arkansas students meeting or exceeding expectations on the ACT Aspire, by content area, 2016 to 2019.2019 ACT Aspire state level


Is anyone improving?

Although statewide scores have not changed, we wanted to highlight some school districts that have demonstrated consistent growth across subjects over the past years. Table 1 highlights these districts, the percentage of students who met or exceeded grade-level expectations in 2018-19, and the percentage point increase in the percentage meeting expectations since the 2015-16 school year.

Table 1. Districts demonstrating consistent increases in the percentage of students meeting or exceeding expectations on the ACT Aspire, by content area, 2016 to 2019.

ACT Gains

It is interesting to note the percentage of students in the districts participating in the Free/ Reduced Lunch program (FRL- a proxy measure of student poverty).  Generally,  districts serving more economically advantaged students have higher rates of students meeting expectations, but we can see variation even among these districts with consistent improvement in performance.  For example, Booneville, Cross County, and Jessieville all serve student population in which 72% of students are eligible for FRL, but proficiency rates vary between them.


Dig into the Data!

There’s so many different variables to consider as we try to make sense of the most recent ACT Aspire scores.  We know you want to dig into the data yourself and see how your district or school is doing so we have created interactive data visualizations. Maps are available for both district and school-level, and you can select specific districts/schools and see how they scored and whether they have improved performance over last year.  There are subject-specific maps available (tabs above the main map) if you want to see one content area in particular. You can also use filters to find schools/ districts with similar rates of economically disadvantaged students enrolled and serving students in similar grade levels.

2019 ACT Aspire school level

In these maps, we use the OEP GPA as the overall indicator of performance. The OEP’s GPA is a weighted measure of student performance that gives the most credit to students who have exceeded expectations and the least credit to those that are in need of support. In this GPA measure, we treat the ACT Aspire test scores similar to the familiar grade point average for individual students: 1.0 is the lowest score, indicating that all students in a districts were In Need of Support, while 4.0 is the highest score, indicating that all students in a districts were Exceeding Expectations on the ACT Aspire. OEP’s GPA is more sensitive to changes in student performance than the % Met Expectations value, and although calculated slightly differently, OEP’s GPA is highly correlated with the ESSA Weighted Achievement measure (r=0.98 in 2018), and the overall ESSA index score (r=0.96 in 2018).

As you can see in the maps above, schools in the upper left hand corner of the state are more likely to be blue, indicating a higher OEP GPA. This is not surprising since we are showing performance on the ACT Aspire, which is correlated with the percentage of students in the school participating in the Free/ Reduced Lunch program (a proxy measure of student poverty), so more economically advantaged areas of the state generally have higher scores.


Digging Deeper Into ACT Aspire

We now turn to examining each content area by grade level.

While overall ACT Aspire results did not change, there were some changes by grade level within each content area. Although we prefer the OEP GPA,  here we use the percent meeting or exceeding benchmarks so we can compare Arkansas performance with national results. When interpreting National Averages, however, it is important to remember that Arkansas students make up an ever increasing percentage of the “national” three – year rolling average.  This is because Arkansas is the only state currently administering the ACT Aspire in grades 3-10.  Alabama and South Carolina, the other states with significant representation in the norming sample, stopped administering the ACT Aspire in 2017 and 2015, respectively.


Mathematics

Figure 2. Percentage of Arkansas students meeting or exceeding expectations on the ACT Aspire Math, by grade, 2016 to 2019.

2019 ACT Math_Grade

Math: We see consistent increases in the percentage of 3rd, 8th, 9th, and 10th graders meeting or exceeding expectations. The increases in 3rd grade performance are important as they might foreshadow higher proficiency levels in later grades (although that pattern isn’t evidenced so far). The continued improvement in 8th-10th grades is great news, since these grades returned the lowest math proficiency rates in the first year of testing.


Science

Figure 3. Percentage of Arkansas students meeting or exceeding expectations on the ACT Aspire Science, by grade, 2016 to 2019.

2019 ACT Science_Grade

Science: Performance in science was generally consistent with prior performance, although there was a slight increase in 9th grade proficiency.  The proficiency of 6th grade had been declining, but held steady this year.


English

Figure 4. Percentage of Arkansas students meeting or exceeding expectations on the ACT Aspire English, by grade, 2016 to 2019.

2019 ACT English_Grade

English: Performance in English was generally consistent with prior performance, although there were slight declines in some grades.

 


Reading

Figure 5. Percentage of Arkansas students meeting or exceeding expectations on the ACT Aspire Reading, by grade, 2016 to 2019.

2019 ACT Reading_Grade

Reading: We can see consistent increases in the reading proficiency in grades 3-5, although these gains are offset by performance declines in the high school grades.

We got to wondering about the performance of the Outstanding R.I.S.E schools from 2018, and if their training in and successful implementation of the R.I.S.E strategies would be evidence in increased reading proficiency for the third graders assessed in 2019. The 3rd grade students tested last spring would have been 2nd graders in the first year of R.I.S.E Arkansas. Figure 6 displays the 3rd grade reading performance of these R.I.S.E. schools and the state overall since ACT Aspire testing began in 2015-16.

Figure 6. OEP GPA for the ACT Aspire 3rd grade reading, 2016 to 2019.

RISE Reading

In 2015-16, schools that would be identified in 2018 as Outstanding R.I.S.E. schools had the same 3rd grade reading performance as 3rd graders across the state as a whole.  This is great because it shows that these RISE schools were scoring no better or worse that the average school across the state. The reading performance in the schools that would be identified as R.I.S.E schools in 2018, began to decline in 2016-17 and 2017-18, while the statewide 3rd grade reading performance continued to increase slightly. In 2018-19, however, there was improved reading performance among 3rd graders in the R.I.S.E schools.  Although we can’t say for sure that the recent increase in 3rd grade reading performance is due to the implementation of R.I.S.E strategies, it is a positive sign that there is improvement in these schools!


So- what are the big takeaways from the preliminary 2019 ACT Aspire results? 

  • Performance in all subject areas is generally the same as last year (and the year before..), although some districts are demonstrating consistent improvement.
  • Performance on the ACT Aspire is related to school/ district poverty rates.
  • OEP GPA is highly correlated with Weighted Achievement and ESSA Index scores.
  • You can use the data resources from OEP to see overall district and school values.
  • Data visualizations can help us see statewide patterns in performance and compare performance to other schools/ districts of interest.

We are happy to be able to share these resources with you and looking forward to seeing the Growth Scores (our favorite!) on the ESSA reports in October. Stay tuned to OEP for more info, and let us know if there’s something you want us to check out!