WISELI is extremely concerned about protecting the identities of Study of Faculty Worklife survey respondents. Since our first survey in 2003, we have taken many precautions to ensure respondent identities are protected, while maintaining a database that is longitudinal, flexible, and useful for answering organizational and research questions into the future.
In order to protect respondent identity, and simultaneously allow for the Study of Faculty Worklife survey data to act as a rich tool for answering important research questions, survey responses are confidential, but not anonymous.
Retaining the ability to match a respondent’s name to his or her survey answers carries an element of risk that the identifiers could be made public. To minimize this risk:
- Survey responses are attached to a random, non–identifying ID number generated by the University of Wisconsin Survey Center (UWSC). Names and identifying personnel numbers are NOT included in the file containing survey response data.
- One file that matches the random ID in the datafile to an identifying personnel ID number is created by the UWSC and provided to WISELI researcher Jennifer Sheridan. Dr. Sheridan stores this identifying file on a secure, password–protected disk. No other WISELI researchers have access to this file, and no one outside of Dr. Sheridan and the UWSC have access to this file.
Retaining the ability to match a respondent’s personnel identification number to his or her survey responses allows researchers to:
- Perform in–depth response analysis to determine whether respondents are representative of all faculty and staff in the sample.
- Link responses to Study of Faculty Worklife surveys over time, in order to perform longitudinal analyses.
- Link responses to participation in WISELI activities such as grant programs and workshops, in order to perform evaluations of WISELI programming.
- Link responses to UW–Madison (publicly available) administrative data data such as attrition, salary, grants, or lab space, to answer relevant research and policy–related questions.
- Link responses to other publicly-available data such as publication record or patent filings, to answer research questions about the relationship between department climate and faculty productivity.
- Retain variables such as department, gender, race and ethnicity, sexual orientation, and citizenship status, which are invaluable to answering questions of equity among various demographic groups. This allows researchers the ability to examine these variables while controlling for respondents’ department or discipline. These variables in combination could be identifying, so WISELI instituted further protections outlined below.
- Maximize response rates by sending out multiple surveys to non–responders, while avoiding confusion and irritation among those who have already responded.
Because faculty and staff at UW–Madison work in highly identifiable environments, some combination of variables on the Study of Faculty Worklife survey could be identifying (for example, the combination of department and gender, or school and race/ethnicity). Because of this possibility, WISELI has pledged that NO RESULTS WILL EVER BE REPORTED AT THE DEPARTMENT LEVEL OR BELOW. Results will only be reported at the divisional, school or college, or other aggregated level that allows for meaningful comparisons between groups but not identification of individuals.
Beginning in 2003, WISELI researchers have allowed for the possibility that researchers outside of the WISELI research group, as approved by a Human Subjects Institutional Review Board (IRB), might want to analyze the Study of Faculty Worklife data. WISELI has developed the following protocols for sharing Study of Faculty Worklife data, in order to minimize the risk of respondent identification.
- All research studies using the Study of Faculty Worklife data must be approved by an IRB and also by the Executive Committee of WISELI.
- No outside researcher will be given access to the file linking individual identifiers to respondent data.
- No outside researcher will be given access to the entire data file. IRB–approved researchers will only be provided extracts of data that include limited variables approved by an IRB.
- All researchers must abide by the agreement to never display results at the department level or below, and must agree to always aggregate data so that individuals cannot be identified.
- Unless specifically approved by a human subjects review board and the executive committee of WISELI, the following variables will not be provided in extracts of the data:
- Indicator of multiple appointments
- Names of schools attended
- Unless specifically approved by a human subjects review board and the executive committee of WISELI, the following variables will always be condensed into general categories and/or will be disguised, and the original variables will not appear in an extract:
- Department—may be provided if identified by a random number, or if aggregated to a broad disciplinary level CIP code
- Race/Ethnicity—will become a White/NonWhite dichotomy
- Employment History—will be represented by calculated variables of time since PhD, time to tenure, and time to promotion
- Age—will be represented in 4 or 5 age brackets
- Salary—a random amount between –$500 and +$500 will be added to salary data
- Other potentially identifying variables—top-coding or aggregation of any variable that might make a person identifiable
- Other protections will be implemented as needed.
These procedures were originally approved by the UW–Madison Social & Behavioral Sciences Institutional Review Board (protocol number #06-01-25) in February, 2003. They have been updated and approved under protocols #SE-2007-0239, #SE-2012-0021, #2015-1601, and #2018-1566.