National Center for Education Statistics

Challenges, changes, and current practices for measuring student socioeconomic status

By Lauren Musu-Gillette

There is an abundance of data and research that shows a relationship between a student’s socioeconomic status (SES) and their academic outcomes. For example, students from low-SES families are far more likely to drop out and far less likely to complete a bachelor’s degree than their peers from middle- and high-SES families.

As we seek to better interpret and understand these and other findings related to student progress, it important for NCES to try to collect accurate and complete measures of student SES.

Percentage distribution of highest level of educational attainment of spring 2002 high school sophomores in 2012, by socioeconomic status (SES)

SOURCE: U.S. Department of Education, National Center for Education Statistics, Education Longitudinal Study of 2002 (ELS:2002), Base Year and Third Follow-up. See Digest of Education Statistics 2014, table 104.91.

Measures of SES usually combine several different statistics, most commonly family income/wealth, parent educational attainment, and parent occupation.[1] In some surveys, NCES is able to collect data directly from parents in order to measure all these component of SES. However, in many assessments and some surveys, NCES is unable to collect this information directly from parents making it difficult to create a consistent measure of SES across the Center.

NCES staff recognizes both the importance of collecting valid and reliable SES data, and the challenges associated with doing so. For example, between 2010 and 2012, NCES convened a panel of experts in the fields of economics, education, statistics, human development, and sociology who provided information on SES, including theoretical foundations, common components, data collection and measurement approaches, and possible implications of a new measure of SES for the National Assessment of Educational Progress (NAEP). There are several challenges for NAEP when considering the inclusion of survey items that can be used to measure student SES. Since NAEP does not include a parent survey, student or school-level data is currently the only potential source for data. However, data on SES can be difficult to collect directly from students as many are unable to accurately respond to questions about their family income or the highest level of their parent or parents’ education.

In terms of school-level data, student eligibility for free and reduced price meals has historically been an important indicator of household income. However, recent changes in the way schools are required to record eligibility for free and reduced price meals has required researchers to reconsider the use of this data point  as a measure of family income, or as a proxy, more generally, for SES. A recent NAEP blog on this topic provides additional information on NAEP-specific considerations, but these changes impact data collection efforts across the agency. Additionally, the free and reduced price meals data only reflect income, which is only part of a complete SES measure, and does not differentiate between middle and high SES students.

Given these changes, NCES is working to  identify other variables that could serve as more reliable and valid measures of student SES.  For example, several NCES staff members are involved with the Alternative SES Measure Working Group as part of the National Forum on Education Statistics. This group recently released the Forum Guide to Alternative Measures of Socioeconomic Status in Education Data Systems. This publication presents advantages and disadvantages for eight alternative measures of SES.  These resources are intended to serve as reference tools for education agencies engaged in identifying, evaluating, or implementing alternative SES measures. They are not data collection instruments and do not represent federal reporting requirements.

Collecting data on and examining differences in educational outcomes by student SES is important to both researchers and educators. As data systems evolve and measures of SES change over time, NCES is committed to researching and collecting the best data possible with the resources available.


[1] For most NCES surveys, parent educational attainment and parent occupation is based on the highest level achieved by either parent and/or guardian in the household.

Financing education: National, state, and local funding and spending for public schools in 2013

By Lauren Musu-Gillette and Stephen Cornman

Spending on public education continues to fluctuate significantly among states and school districts, according to two NCES reports released Wednesday. The reports also show that, nationally, spending on elementary and secondary education declined for the fourth straight year.

The two First Look reports, Revenues and Expenditures for Public Elementary and Secondary Education: School Year 2012–13 (Fiscal Year 2013) and Revenues and Expenditures for Public Elementary and Secondary School Districts: School Year 2012–13 (Fiscal Year 2013), are based on data from the National Public Education Finance Survey (NPEFS), a component of the Common Core of Data (CCD).

In 2013, expenditures for public elementary and secondary education totaled $606.5 billion, which translates to $10,763 in per student spending[i] on a national level. Public elementary and secondary school finance can vary considerably depending on the state or school district.

At the state level, spending per student ranged from a low of $6,432 in Utah to $20,530 in the District of Columbia (D.C.). After D.C., per student spending was next highest in:

  • New York ($19,529);
  • New Jersey ($18,523);
  • Alaska ($18,217);
  • Connecticut ($17,321); and
  • Vermont ($17,286).

Current expenditures per pupil for public and secondary education, by state: Fiscal year 2013

SOURCE: U.S. Department of Education, National Center for Education Statistics, Common Core of Data (CCD), “National Public Education Financial Survey

Among the 100 largest school districts in the nation, those with the highest spending per student were:

  • New York City School District ($20,331);
  • Boston City Schools, Massachusetts ($19,066);
  • Philadelphia School District, Pennsylvania ($16,381);
  • Anchorage School District, Alaska ($15,391);
  • Montgomery County Public Schools, Maryland ($15,080); and
  • Baltimore City Schools, Maryland ($15,050).

As a nation, we spend more per-student on elementary and secondary public education than we did 10 years ago, but 2013 represents the fourth straight year that our national per-student spending has fallen.  In order to compare spending from one year to the next, expenditures are converted to constant dollars, which adjusts figures for inflation. From 2002–03 to 2012–13, spending per student enrolled in public elementary and secondary schools in the U.S. increased by 5 percent (from $10,455 to $11,011 in constant 2014–15 dollars). Spending per student increased at least 1 percent per year between 2003–04 and 2007–08, and peaked in 2008–09 at $11,621. It has decreased each year since then, with the greatest decrease occurring from 2008–09 to 2011–12.  

Current spending per student in fall enrollment in public elementary and secondary schools: 2002–03 through 2012–13

NOTE: Spending is reported in constant 2014–15 dollars, based on the Consumer Price Index (CPI).
SOURCE: U.S. Department of Education, National Center for Education Statistics, Common Core of Data (CCD), "National Public Education Financial Survey," 2002–03 through 2012–13. See Digest of Education Statistics 2015, table 236.65.

The recently released reports also present national and state level data on public school funding[ii] by source. Total funding per pupil decreased by 1.2 percent on a national basis and decreased by 1 percent or more in 26 states from 2012 to 2013, after adjusting for inflation. The 50 states and D.C. reported $603.7 billion in funding collected for public elementary and secondary education in 2013. State and local governments provided $547.8 billion, or 91 percent of all funding; and the federal government contributed $55.9 billion or 9 percent of all funding.

The percentage of total funding from federal sources accounted for 9 percent of total funding in both 2002–03 and 2012–13; however, there were notable fluctuations during this period. The federal percentage increased from 8 percent of funding 2007–08 to 13 percent of funding in 2010–11. This increase reflects the impact of the American Recovery and Reinvestment Act (ARRA). As the funds from the program were spent, the federal percentage decreased to 10 percent of total funding in 2011–12 and to 9 percent in 2012–13. Local sources accounted for 46 percent of total funding in 2012–13, the highest percentage in the past 10 years. The percentage of total funding from state sources decreased from 49 percent in school year 2002–03 to 45 percent in school year 2012–13.

Percentage of funding for public elementary and secondary schools, by source of funds: 2002-03 through 2012-13

SOURCE: U.S. Department of Education, National Center for Education Statistics, Common Core of Data (CCD), "National Public Education Financial Survey," 2002–03 through 2012–13. See Digest of Education Statistics 2015, table 235.10.

[i] Spending refers to current expenditures.Current expenditures are comprised of expenditures for the day-to-day operation of schools and school districts for public elementary and secondary education, including expenditures for staff salaries and benefits, supplies, and purchased services. Current expenditures include instruction, instruction-related, support services (e.g., social work, health, and psychological services), and other elementary/secondary current expenditures, but exclude expenditures on capital outlay, other programs, and interest on long-term debt. 

[ii] Funding refers to revenues. Revenues are comprised of all funds received from external sources, net of refunds, and correcting transactions. Noncash transactions, such as receipt of services, commodities, or other receipts in kind are excluded, as are funds received from the issuance of debt, liquidation of investments, and nonroutine sale of property.

Beginning postsecondary students: Persistence and attainment after 3 years

By David A. Richards

The number of students enrolling in postsecondary education has increased over the past several decades. While this increase in enrollment shows that more students are pursuing postsecondary credentials and degrees, it is also important to consider the number of students that go on to complete their postsecondary education. Data from the Beginning Postsecondary Students Longitudinal Study (BPS) can help researchers, policy-makers, educators, and the public answer questions about whether students are persisting through their educations and earning credentials. BPS data can also help answer questions about how these outcomes may differ by institutional and student-level characteristics.

The recently released Persistence and Attainment of 2011-12 First-Time Postsecondary Students After 3 Years contains findings from data collected from the Beginning Postsecondary Students (BPS) longitudinal study. This report answers questions such as, what percentage of first-time students who began postsecondary education in 2012 were still enrolled three years later? How many had earned a credential, and how did rates differ across different types of postsecondary institutions and degree programs?

This first look report on BPS:12/14 data shows that, among 2011–12 first-time postsecondary students, 7 percent had completed a certificate, 7 percent had completed an associate’s degree, and 1 percent had completed a bachelor’s degree at any institution within 3 years. Of the students who had not yet earned a credential, 39 percent were enrolled at a 4-year institution, 16 percent were enrolled at a less-than-4-year institution, and 30 percent were not enrolled at any institution by the spring of 2014.

At the baccalaureate level, among students who first enrolled in 4-year institutions and were seeking bachelor’s degrees, 1 percent had completed a certificate, 1 percent had completed an associate’s degree, and 3 percent had completed a bachelor’s degree at any institution within 3 years. Another 73 percent of the students seeking a bachelor’s degree were enrolled at a 4-year institution, 6 percent were enrolled at a less-than-4-year institution, and 16 percent were no longer enrolled at any institution.

Percentage distribution of first-time public 2-year college students 3 years after entry, by student age: 2012–14

! Interpret data with caution. Estimate is unstable because the standard error represents more than 30 percent but less than 51 percent of the estimate.
NOTE: Includes first-time postsecondary students starting at a Title IV eligible postsecondary institution in the 50 states and the District of Columbia in 2011-12.
SOURCE: U.S. Department of Education, National Center for Education Statistics, 2012/14 Beginning Postsecondary Students Longitudinal Study (BPS:12/14).

BPS also contains data on attainment based on student characteristics such as race, ethnicity, and age. These data can be examined by the type of program in which students are enrolled and the first institution in which they enrolled. For example, among first-time postsecondary students beginning at a 2-year public college, 4 percent of those students who were age 18 or younger when they enrolled in 2011 had completed a certificate by the spring of 2014, 14 percent had completed an associate’s degree, 45 percent were still enrolled in a postsecondary institution, and 37 percent were no longer enrolled in postsecondary education. For students that were age 30 or older when they enrolled, 8 percent had completed a certificate by the spring of 2014, 8 percent had completed an associate’s degree, 29 percent were still enrolled, and 56 percent were no longer enrolled. Differences by other student characteristics, such as sex, race/ethnicity, dependency status, and parental educational attainment are available in the report.

BPS surveys nationally representative cohorts of first-time, beginning students at three points in time: at the end of their first year, and then three and six years after first starting in postsecondary education. It collects data on a variety of topics, including student demographic characteristics, school and work experiences, persistence, transfer, and degree attainment. BPS is a detailed follow-up to the National Postsecondary Student Aid Study (NPSAS), a nationally representative, cross-sectional study of U.S. postsecondary students designed to collect data on how postsecondary students pay for their education. The most recent BPS study, (BPS:12/14) used the 2012 NPSAS data as its base year (which included enrollment characteristics, education aspirations, and demographics) and conducted its first follow-up in 2014. (Another follow-up will be conducted in 2017.) During the 2014 follow-up study, BPS participants  were surveyed on their enrollment patterns since 2012—providing information about transfers, stopouts[1], attendance, and credentials earned—as well as on their employment histories. Study data were also drawn from a variety of other resources.

If you’re interested in comparing these findings to earlier BPS iterations, you can find earlier first look reports on the BPS publication page. BPS:12/14 data are also available for analysis through the online DataLab tool. If you have questions about the report or this data, please reach out to the National Center for Education Statistics at or by phone at (800) 677-6987.


[1] A stopout is a temporary break in enrollment.

Virtual schools: Measuring access to elementary and secondary education in online environments

By Mark Glander

Many people are familiar with the increasing availability of online education at the postsecondary level, but did you know that the number of virtual elementary and secondary schools is also growing? Virtual schools can offer flexibility to students who may have difficulty accessing or attending traditional schools, or as an alternative to homeschooling for parents who elect not to enroll their children in traditional brick and mortar schools. As the number of schools offering virtual education increases, it is important to be able to track these schools.

To gain a better understanding of the role virtual schools play in public elementary and secondary education, NCES added a flag identifying these schools to its Common Core of Data (CCD). The CCD is an annual collection of data from all public schools, public school districts, and state education agencies in the United States. The recently released School Year 2013–14 collection includes the new virtual school flag. For this purpose, a virtual school is defined as, “A public school that offers only instruction in which students and teachers are separated by time and/or location, and interaction occurs via computers and/or telecommunications technologies. A virtual school generally does not have a physical facility that allows students to attend classes on site.”

Twenty-eight states and the District of Columbia reported having one or more virtual schools for a total of 478 virtual schools in the U.S. in 2013–14. Florida reported the most of any state with a total of 182. A new data item is often under-reported in the first year of collection; ten states and other jurisdictions did not report having any virtual schools or reported virtual schools as not applicable (California, Delaware, North Dakota, Texas, Washington, the Department of Defense Education Activity, American Samoa, Guam, Commonwealth of the Northern Mariana Islands, and the U.S. Virgin Islands). 

All but 12 of the reported schools were “regular” schools, meaning they offered a general academic curriculum rather than one focused on special needs or vocational education. 

The CCD distinguishes several types of local education agencies, defined by their level of governance.  Almost all virtual schools were administered by regular, local school districts (350 schools). Most other virtual schools were administered by independent charter school districts (116 schools), which are districts composed exclusively of charter schools.

The two states with the largest number of students in virtual schools were Ohio (38,169) and Pennsylvania (36,596).  Idaho had the largest percentage of students in virtual schools (2.4 percent), followed by Ohio (2.2 percent), and Pennsylvania (2.1 percent).

CCD identifies four school levels:  primary, middle, high, and “other”.  “Other” includes schools that span these categories and schools with high school grades but no 12th grade. A total of 309 of the 478 virtual schools had a school level of "other".  These schools accounted for 84 percent of students in virtual schools.

To see tables summarizing the above data, please visit our web page –

To learn more about the CCD, please see our latest report, or visit our web page.  You can also access CCD data files for additional information about public elementary and secondary schools.  

Exploring a range of educational outcomes within and across countries: Sub-national data supplement to Education at a Glance 2015

By Lauren Musu-Gillette and Tom Snyder

Situating educational and economic outcomes in the United States within a global context can help researchers, policy makers, and the public understand how individuals in the U.S. compare to their peers internationally. The annual publication Education at a Glance produced by the Organisation for Economic Co-operation and Development (OECD) provides information on the state of education in many countries across the world. While these data are instrumental in helping us to understand how the U.S. compares to other OECD and partner countries on a number of key educational and economic outcomes, national averages can mask the high degree of variation that can occur within individual countries. In order to address this, several OECD and partner countries, including the U.S., provided sub-national data on several select indicators previously only available at the country level.   

These data, posted on the NCES website, serve as a supplement to Education at a Glance 2015 and provide select sub-national data for six indicators in this edition. These include data on educational attainment by selected age groups, employment rates by educational attainment, annual expenditure per student, enrollment rates by age, enrollment rates in early childhood and primary education, and enrollment rates and work status of 15-29 year-olds.

In order to understand the amount of variability in an indicator for a particular country, we can compute a ratio of the state, territory, or region with the highest percentage to that with the lowest percentage on any given metric. For example, within the United States, the percentage of 25- to 34-year-olds who completed any level of postsecondary education in 2013 ranged among the states from 30 percent in Nevada to 55 percent in Massachusetts.

The ratio of high to low percentages of 25- to 34-year-olds completing postsecondary education in the United States (2.4) was among the largest of the reporting countries.  The ratio was higher in Brazil (5.8) with a range of 6 to 31 percent, and in Spain (2.8), with a range from 21 to 58 percent. The U.S. ratio was slightly higher than in Canada and Russia (both 2.3).  The ratio was lower in Sweden (1.8) and lowest in Slovenia (1.0), Ireland (1.2), and Belgium (1.2). The high to low ratio between OECD countries was 2.8, ranging from a low of 24 percent to a high of 68 percent.

Average percentage of the 25-34 year old population with postsecondary education (with subnational high/low value) in selected OECD and partner countries: 2014

NOTE: Countries are ranked in ascending order of the average percentage of the 25-34 year old population with postsecondary education. Data years differ. Data for Canada is from 2012, while data for the United States and Brazil is from 2013. Data for all other countries is from 2014.

SOURCE: OECD. Table A1.3a. See Annex 3 for notes and sub-national Summary Table A1.3a.

Regional policy makers can benefit most from the comparisons presented in Education at a Glance when they can compare the results from their own sub-national areas with national and sub-national data from other countries. It is not surprising that large federal countries, such as Canada, Germany, and the United States, in which education is largely controlled by regional authorities, might have large internal variations. But, many other countries with centralized education systems such as Spain and Sweden have substantial variations within their countries as well. These new sub-national data can help illuminate these differences and provide additional information to U.S. states on how they compare to their peers both within the U.S. and internationally.