The 2012 National Survey of Science and Mathematics Education (NSSME) was designed to provide up-to-date data to examine status and trends in the areas of teacher background and experience, curriculum and instruction, and the availability and use of instructional resources. A total of 7,752 science and mathematics teachers and 1,504 schools across the United States participated in this survey. As part of dissemination, public release datasets from the 2012 NSSME data are available for secondary analysis.
The 2012 National Survey’s sampling involved stratification, clustering, and unequal probabilities of selection, all of which must be reflected in standard error calculations so that the results of analyses are representative of all teachers in the U.S. Based on the sample design for this study, a set of 75 jackknife 2 replicate weights was created for calculating standard errors for school and teacher estimates.
For access to the data, complete the form linked below (all fields are required) and fax it to Horizon Research, Inc. (919-493-7589) or scan and email the completed form to <nssme@horizon-research.com>.
The form below is a PDF file which can be read by Adobe Acrobat Reader (if you do not already have Acrobat Reader, you can download it for free from Adobe’s Website).
2012 NSSME Public Release Data Terms of Use (PDF, 10KB)
After receiving the completed form, we will contact you via email with instructions for accessing the data. The data are available in either SPSS v22 or tab-delimited text format. Accompanying the datasets will be:
- An HTML data dictionary for each dataset
- The 2012 NSSME instruments
- The NSSME Public Release Datasets User Manual
- Tutorials on how to analyze NSSME data using Wesvar.
NOTE: Although use of the weights in the public release dataset will provide correct parameter estimates in programs like SPSS, standard errors will be substantially underestimated, greatly increasing the likelihood of “false positives” from statistical tests. A number of methods exist for adjusting standard errors to account for sampling weights. One method is to use a statistical program designed for analyzing data from complex samples, such as WesVar, SUDAAN, Stata, or the survey procedures in SAS v9.