Skip Navigation


The HSLS:09 project team is grateful for researchers’ interest in analyzing the HSLS:09 dataset. Many users send similar questions about data access and content. This page provides answers to such Frequently Asked Questions and will be updated as additional questions emerge.

Are the HSLS:09 data released yet?
Yes, the base year data were released in August 2011. Data from the first follow-up (collected in 2012) were released in November 2013. Data from the 2013 Update and High School Transcript Collection were released in June 2015.

Is there a restricted-use dataset?
Yes, there are both public-use and restricted-use datasets available to data users.

How do the public-use and restricted-use data differ?
The public-use data files reflect alteration or suppression of some of the original data. Such edits minimize the risk of disclosing the identity of responding schools and the individuals within them. The restricted-use data provide school information, most of which had to be suppressed on the public-use dataset, along with more student-level variables with less alteration or suppression.

Which data should I get – public-use or restricted-use?
It depends on your variables of interest! Review the complete list of variables on the EDAT. Then determine from those variables’ frequencies whether any of the values are suppressed, thus requiring you to analyze the restricted-use data. Appendix L in the Data File Documentation compares public-use data with restricted-use data and presents which variables have been altered or suppressed and how:

How do I get the restricted-use data?
You must apply for a restricted-use data license through the NCES website. To apply for this license, you must swear to adhere to all regulations, rules, guidelines, policies, etc. that govern the use of HSLS:09 restricted-use data.

Whom do I contact for the restricted-use data?
The HSLS:09 project team and project officer are not involved with the restricted-use data license application process. This is a separate department at NCES (

How long does it take to get a restricted-use data license?
Typically, three weeks, provided there is no problem with your application.

What if my project changes?
Licenses can be amended to obtain additional HSLS:09 restricted-use data or restricted-use data from other NCES surveys.

How can I access the public-use data?
Access the public-use data, quickly and easily, from our website, here:

You can tag which variables you would like to download from the vast HSLS:09 public-use dataset by using the EDAT, a data analysis tool operated on the NCES website. Agree to the terms of EDAT usage, select HSLS:2009/13 when asked, and explore the variables and frequency distributions of the public-use data. Once you choose the variables you wish to download, select in which statistical software package format you would like to download the data, and follow the on-screen directions. The data will be downloaded to your hard drive almost instantaneously.

I’ve downloaded the data on the EDAT, but it’s not working. Why?
You must edit the syntax file you download for your specific statistical software package to indicate where on your hard drive you want the data to be stored. NCES cannot predict the name of your computer’s hard drive, so we leave that editing to you. You must direct your computer where to load and store the data through revising that syntax file before running it.

This document provides more detailed instructions on using the EDAT to download data:
HSLS EDAT Instructions MS Word (16 KB)
EDAT instructions for MS Word (24 KB)

If you still experience trouble, please contact the HSLS:09 project team. We can help troubleshoot and offer more specific solutions.

So I have the data. Now what?
First, download the Data File Documentation. This documentation provides critical, valuable information for understanding and analyzing the data. The Data File Documentation also includes a codebook with the frequencies for all variables in the public-use dataset. Read through this document to familiarize yourself with the sample, the data structure, the research questions which the variables are designed to address, the potential for addressing other research questions, and so much more.

Then, explore! Examine the frequency distributions of your variables of interest, start running analyses to learn how the variables work together.

What are weights? Should I use them?
Analytic weights are used in combination with software that accounts for HSLS:09’s complex survey design to produce estimates for the target population, with appropriate standard errors. In addition, because of the comparatively low unit response rates for parents and teachers, special student weights—adjusted for parent, mathematics teacher, and science teacher nonresponse—were also produced. Five sets of analytic weights were computed for HSLS:09:

  1. a school-level weight that can be used for school-level analyses involving the school administrator and school counselor questionnaires;
  2. a student-level weight for student-level analyses;
  3. a parent weight for student-level analyses involving the parent questionnaire;
  4. a math teacher weight for student-level analyses with the math teacher questionnaire data; and
  5. a science teacher weight for student-level analyses with the science teacher questionnaire data. The parent and teacher weights presuppose that parents and teachers provide contextual data for participating students, and that the student is the unit of analysis.

Variance estimation is provided through two means: BRR (Balanced Repeated Replication) provided on both public- and restricted-use files and a Taylor series linearization (available on the restricted-use file). The BRR approach to calculating HSLS:09 standard errors is recommended, although both methods give similar results.

More information is available from the project officer or the HSLS:09 project team at

Why do I get only 21,444 cases when you say the sample includes >24,000 students?
The Basic Documentation has a discussion in the Executive Summary of the process by which we statistically adjust for the fact that 21,444 students responded out of the >24,000 drawn as a representative sample. The documentation also discusses the need to incorporate the weights we provide in the dataset to make that adjustment in your analysis.