MIT Institutional Research
MIT-Cleaning-Air-3-press.jpg

Administrative Analysis

Record level survey files can be requested by offices within MIT for administrative analysis.

Our policy on sharing record level survey data for administrative use

In order to maximize the use of data collected in institute surveys, Institutional Research (IR) encourages other offices at MIT to leverage existing survey datasets when possible rather than conducting additional surveys. Because IR has the responsibility for the protection of individual respondents, IR has established policies for the sharing of datasets. 

Administrative analysis is generally conducted to inform internal practices, policies, or decisions. Assessment is one form of administrative analysis. Administrative analysis does not attempt to extrapolate to populations beyond the Institute. While it might be informed by theory, the goal is to inform administrators, not develop or test academic theories. 


Record level survey files can be requested by offices within MIT for administrative analysis

Survey analysis is, and should continue to be, conducted by analysts in IR who are highly experienced and are intimately familiar with the data. At times, though, content experts in administrative units may require direct access to record-level survey data to conduct additional analysis. 

All individuals who will have access to a survey data file must be specified in the request and must have completed Human Subjects training. By receiving a survey data file, recipients agree to:

  1. not re-identify a respondent

  2. not report data in any way that might allow a respondent to be identified

  3. to secure the data in transit and at rest

  4. not provide the data file in whole or in part to a third party

Record level survey files will include a unique identifier, but not respondents’ MITID, MIT krb, name, or email address. Institutional research recommends that any analysis of the data use a minimum row or column total of 15 and a statistical significance threshold of .001.

Requests for survey datasets should be sent to the Director of Institutional Research.

October 10, 2018