Student Achievement in the US-Mexico Border: Analyzing the 2013 NAEP Eighth Grade Mathematics Assessment

Transcription

1 Student Achievement in the US-Mexico Border: Analyzing the 2013 NAEP Eighth Grade Mathematics Assessment Study Objectives The purpose of this study is to explore factors associated with student achievement in the US-Mexico Border region. Specifically, this study aims to answer the following research questions: 1) What are the characteristics of students in the US-Mexico border among key variables? 2) How are the students in the US-Mexico border performing in comparison to the rest of the nation? 3) What are the factors that may explain performance differences? Introduction and Background According to the National Center of Health Statistics (Notzon, 2009), the US-Mexico border region is defined as 44 U.S. counties within 62 miles (100 km) of the border. This study focuses on the educational conditions and outcomes of students living in these counties. For the purposes of this paper, the term border region shares the same definition put forward by the NCHS, and does not include residents of the Mexican side of the border. In the United States, four states share the Mexican border: California, Arizona, New Mexico and Texas. As of the 2010 Census, the total population of these states is 70.9 million people. Of those, 7.6 million people live within the border region. Initially, the population increase between 1990 and 2000 was beneficial to the area. However, conditions such as environmental degradation, health care and housing worsened each year after 2000 (La Frontera, 2004). Currently, the border region s economy is characterized by low-wage, low skill jobs. These conditions have made it hard for parents to provide sufficient resources and support for their children. Recent research gives a paradoxical picture on academic performance for border students. One study has shown that student achievement within the border region has lagged behind the rest of the nation (La Frontera, 2004). Another study shows that students who were born in Mexico that immigrated to the U.S. had higher grades than third-generation Latino students within the U.S. (Academic Performance, 2001). It is, therefore, imperative to investigate students demographics and achievement to identify factors that may explain difference in performance between border students and the rest of the nation. To capture these factors and outcomes in the US-Mexico border region, county-level data from the U.S. Census Bureau was used in conjunction with data from the National Assessment of Educational Progress (NAEP). In the past, data from NAEP have traditionally been used to identify educational trends and characteristics for individual states or the nation as a whole. This type of non-state regional analysis is an innovative use of the NAEP national sample.

2 Data and Methods This study uses data from the NAEP mathematics assessment administered to eighth graders in To account for the design effects of NAEP s complex sampling strategy, this study estimated variance using the Jackknife repeated replication technique. errors were calculated using STATA survey module and the appropriate weighting and design variables included in the NAEP data set. Since there is no geographic identifier in NAEP that would allow one to directly compare US-Mexico counties to the rest of the nation, a border flag was created from the county definitions provided by the NCHS. We identified the county FIPS code of each NAEP school using the 2013 Common Core of Data (CCD) and then flagged any schools located in a border county. NAEP schools were then linked to information on county-level economic status from the American Community Survey (ACS) of the U.S. Census Bureau (2011). Because NAEP is traditionally designed to describe educational trends and characteristics for individual states or the nation as a whole, not for individual regions of the country, sampling weights for student in Appalachia have been adjusted to the region s population control totals using a three-way raking procedure. The first and second research questions are answered by examining significant differences in the proportions of students assessed by subcategories, and differences in average NAEP scale scores between student populations. The third research question is addressed using OLS regression with average scale scores as the dependent variable. Results Descriptive Statistics The composition of US-Mexico border students is rather different than the rest of the nation on multiple variables (Table 1). Most notably, almost three quarters (72%) of all US-Mexico border students are Hispanic. This is significantly different from the rest of the nation, of which only 22% of students are Hispanic. Students in the border region are also predominantly (66%) eligible for the National School Lunch Program (NSLP). Additionally, 13% of students are English Language Learners (ELL), which is significantly higher than the rest of the nation (5%). All of these variables are consistent with the current literature, as the US-Mexico border is known for having a significant amount of students whose parents do not speak English at home, where some families do not have high-paying jobs and are predominately of Hispanic descent. When disaggregating the percent distribution by school characteristics, we see similar results (Table 2). More than 6 out of 10 (63%) students in the border region attend schools that have more than 75 percent non-white enrollment. Additionally, about two thirds (67%) of students attend schools that are more than 75 percent NSLP-eligible. This shows that the majority of students attending schools in the US-Mexico border region are predominately non-white students who also are NSLP eligible. This is significantly different than the rest of the nation, where only 24% of student attend schools that are more than 75 percent in either NSLP-eligible enrollment or percent non-white enrollment. Performance

3 There were several performance differences between the populations of students when comparing the border region to the rest of the nation for both student characteristics and school characteristics (Tables 3-4). For example, Black students scored significantly higher in the border region (271.1) compared to black students in the rest of the nation (262.9). Within the border region, students who did not have a disability had an average score of and students with a disability had a score of Both scores were lower than similar students in the rest of nation who scored and 248.6, respectively. Additionally, male border students scored lower (276.7) than males in the rest of the nation (284.2), whereas female students did not have a significant difference between the two populations. There were no significant differences in scores between border region students and students in the rest of the nation for any categories of Parental Education, Eligibility for NSLP and ELL status. Table 4 also shows important performance differences between the two regions. Border region students scored higher than their counterparts when categorized by several school characteristics. For example, border region students scored higher in percent non-white enrollment for the Up to 25% and More than 75% categories than similar students in the rest of the nation. Similarly, border region students scored higher in percent NSLP-Eligible enrollment in the From 50 to 75% and More than 75% categories compared to their counterparts. Lastly, border region students scored lower in Suburban and Town schools than similar students in the rest of the nation. Regression Analysis To determine the factors associated with the achievement of students in the border region, a regression analysis of eighth grade NAEP mathematics scores on selected characteristics was conducted for students in the border region and in the rest of the country (Table 5). Regression results indicated that a baseline student a White male living in an suburban locale, who is not eligible for NSLP, attends a non-charter public school, is not designated as ELL or having a disability, and whose parent graduated from college scored approximately 10.4 points lower in the border region compared to the rest of the nation. The R- squared estimates for both models are approximately.35, indicating that about 35 percent of the variation in students performance can be explained by the linear regression models. Overall, many of the variables included in the regression model show a negative effect on the performance of students in both the border region and the rest of the nation, after controlling for all other variables. In the border region regression, many variables did not have a significant effect on the performance of students, whereas all the variables in the rest of the nation regression model had a substantial effect on the performance of students. After controlling for other variables, attending a charter had a positive and substantial effect on the performance of students in both the border region and the rest of the nation (14.5 and 3.8, respectively). This result could be a product of the fact that in school year , California enrolled the largest number of students in charter schools and Arizona had the third highest percentage of charter school enrollment as a percentage of total public school enrollment (Condition of Education, 2015). Additionally, after controlling for other variables, being of Asian race/ethnicity have a positive and substantial effect on the performance of students in both the border region and the rest of the nation (21.58 and 14.31, respectively). Discussion and implications for policy

4 Throughout the report, there are several instances where the data displays results that are not regularly discussed in the literature. It is unknown why in the border region Black students outperform their counterparts in the rest of the nation. Similarly, it is interesting to note that students in the border region outperform their counterparts in select subcategories for percent non-white enrollment. Both of these results are interesting to researchers, especially since there is a large proportion of students in the More than 75 percent non-white enrollment category. It is therefore important for researchers and policymakers to investigate why this relationship exists. Another key finding is the significance in the charter-school variable within the border regression model. While on representing 3% of students in the region, it is noteworthy that attending a charter school was associated with a nearly 15-point increase in student performance, after controlling for other variables in the model. It may be that charter schools, if expanded, could contribute an important element to academic performance for students in the border region. Policymakers and researchers can benefit from this knowledge and should further explore the impact and importance that charter schools may have on the region. Suggestions for future research There were several limitations to the study. Since the researchers were primarily using NAEP data, the data is necessarily a snapshot of student performance. It would be beneficial if additional research used other longitudinal data over a period of time. Another limitation to using NAEP data is the lack of variables that focus on the specific issues that are present in the literature. One common theme for students within the border region is whether or not the student was born in the U.S. This would allow researchers to investigate if a student enrolls in multiple schools, over multiple locations throughout the child s elementary school years. NAEP does not have these specific types of variables available. There are many ways to compare students within a region to students outside of that region. Since our approach is comparing border students to the rest of the nation (all students within the U.S. that are not flagged as border students), other researchers could reassess what group of students should be the comparison group to border students. One suggestion could be assessing how students within the same state (but not identified as border students) compare to students within the border, as opposed to the entire student population outside of the border.

5 References Kena, G., Musu-Gillette, L., Robinson, J., Wang, X., Rathbun, A., Zhang, J., Wilkinson-Flicker, S., Barmer, A., and Dunlop Velez, E. (2015). The Condition of Education 2015 (NCES ). U.S. Department of Education, National Center for Education Statistics. Washington, DC. Retrieved [11/1/15] from McRobbie, J., & Villegas, M. (2004). La Frontera: Challenges and Opportunities for Improving Education Along the US-Mexico Border. WestED. Notzon, S. (2009). Gaps in border health research. U.S.-México Border Health Commission. Second Binational Border Health Research Forum. La Jolla, California. Padilla, A. M., & Gonzalez, R. (2001). Academic performance of immigrant and US-born Mexican heritage students: Effects of schooling in Mexico and bilingual/english language instruction. American Educational Research Journal,38(3),

6 Figure 1. Counties in the US-Mexico Border Region Source: Sam Notzon, National Center for Health Statistics,

7 Figure 2. Border region as defined by the La Paz agreement of 1984 Source:

8 Table 1. Percentage distribution of 8 th grade public school students, by U.S.-Mexico Border region and select student characteristics: 2013 Border Region Rest of Nation Percent Percent Proportion of nation Gender Male Female Race/Ethnicity White 18.0 * Black 2.9 * Hispanic 72.3 * Asian/ Pacific Islander 4.4 * American Indian/Alaskan Native 0.8 * Other 1.6 * Eligibility for NSLP Not-Eligible 34.2 * Eligible 65.8 * Parental Education Disability Did not finish H.S * Graduated H.S Some college Graduated college 37.2 * Omitted/Don't know Not S.D * Student with disability 7.0 * ELL status Not ELL 87.0 * ELL 13.0 * * Significantly different from rest of nation at p <.05 SOURCE: U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics, National Assessment of Educational Progress (NAEP), 2013 Eighth Grade Mathematics Assessment.

9 Table 2. Percentage distribution of 8 th grade public school students, by U.S.-Mexico Border and select school characteristics: 2013 Border Region Rest of Nation Public school type Percent Percent Non Charter Charter Percent non-white enrollment Up to 25% From 25 through 50% From 50 to 75% More than 75% 62.8 * Percent NSLP-eligible enrollment Locale Up to 25% From 25 through 50% From 50 to 75% More than 75% 66.9 * Urban 46.0 * Suburban 34.9 * Town 8.3 * Rural 10.7 * Not applicable. * Significantly different from rest of nation at p <.05 SOURCE: U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics, National Assessment of Educational Progress (NAEP), 2013 Eighth Grade Mathematics Assessment. U.S. Census Bureau, 2010 American Community Survey (ACS).

10 Table 3. Average mathematics scale scores of 8 th grade public school students, by U.S.-Mexico Border region and select student characteristics: 2013 Border Region Rest of Nation Estimate Estimate Overall Mean Gender Male * Female Race/Ethnicity White Black * Hispanic Asian/Pacific Islander American Indian/Alaskan Native Other Eligibility for NSLP Non-Eligible Eligible Parental Education Disability Did not finish H.S Graduated H.S Some college Graduated college Don't know Not S.D * Student with disability * ELL status Not ELL ELL * Significantly different from rest of nation at p <.05 SOURCE: U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics, National Assessment of Educational Progress (NAEP), 2013 Eighth Grade Mathematics Assessment.

11 Table 4. Average mathematics scale scores of 8 th grade public school students, by U.S.-Mexico Border region and select school characteristics: 2013 Border Region Rest of Nation Public school type Estimate Estimate Charter Not charter Percent non-white enrollment Up to 25% * From 25 through 50% From 50 to 75% More than 75% * Percent NSLP-eligible enrollment Locale Up to 25% From 25 through 50% From 50 to 75% * More than 75% * Urban Suburban * Town * Rural Not available. Not applicable. * Significantly different from rest of nation at p <.05 SOURCE: U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics, National Assessment of Educational Progress (NAEP), 2013 Eighth Grade Mathematics Assessment. U.S. Census Bureau, 2010 American Community Survey (ACS).

12 Table 5. OLS Regression analysis of mathematics performance of 8 th grade public school students, by U.S.-Mexico Border region: 2013 Border Region Rest of Nation Coefficient Coefficient Intercept * * Female * Black * * Hispanic * Asian 21.6 * * Other * Eligibility for NSLP * * Parental Education Did not finish H.S * * Graduated H.S * * Some college * Don't know * * Urban * Town * Rural -3.5 * * Charter 14.5 * * Student with disability * * ELL * * R Square * Significantly different from intercept at p <.05 SOURCE: U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics, National Assessment of Educational Progress (NAEP), 2013 Eighth Grade Mathematics Assessment.