Prospective COACHE researchers should be aware of both the opportunities and the limitations of our data.
Some of the opportunities:
-
Our dataset includes many variables provided by institutions for all records--respondents and nonrespondents. These include gender, race/ethnicity, academic area, rank, and tenure status, and in some cases, US citizenship, CIP, year of hire, and year of last appointment. These data can be essential for statistical weighting techniques that might be necessary in your analysis.
-
COACHE is a census, not a sample, so we have near-complete representation of every pre-tenure and/or tenured faculty member at participating institutions.
-
Most institutions in our dataset have administered a COACHE survey more than once, offering researchers the capacity for longitudinal analysis at the institutional and, sometimes, unit level.
Some of the limitations:
-
Institutions are not randomly included in COACHE. Their presidents and provosts elect to join the project at a not-insignificant cost. This results in a dataset that is not as a whole representative of faculty profiles across U.S. higher education.
-
Faculty are not randomly selected to participate. All eligible faculty are invited to respond.
-
The data exclude faculty in their first year of employment at an institution and those who are in their terminal year after being denied tenure.
-
COACHE data does not (yet) include part-time/adjunct faculty.
-
We cannot deliver information in our dataset that identifies individual faculty (because of IRB requirements) nor individual institutions (because of contractual agreements with them).
The codebook you receive in the first step of the data request process is not a perfect representation of our dataset. Some variables were not gathered in some years, and institutional participation varied greatly from year to year. For example, not until 2014 did COACHE expand to include community colleges, and only two systems of community colleges have participated.