We recently received a question about how time is handled in our dataset, so we put our upcoming geographic analyses on the back-burner to take a quick look at how time on the job market influences employment patterns. As a reminder, we only have data for each of our grads at one point in time; while this prevents us from determining the specific employment trajectory of a given individual, we can use the full sample to examine the association between graduation year and employment type.
A simple logistic regression tells us, somewhat predictably, that the odds of faculty/professor employment are lower among more recent grads than among grads who have spent more time on the job market. Similarly, the odds of other types of higher education employment (postdoctoral fellowships, research fellowships, etc.) are higher among more recent grads. When we examine this graphically, we find that the associations are most pronounced at the upper and lower ends of our graduation year continuum; the proportion of grads from 2007 to 2010 working in either category is essentially equivalent, as illustrated in the plot below.
The relationship between graduation year and higher education employment is similar across every discipline with sufficient data, and the trends are consistent for both men and women:
We also looked at the relationship between graduation year and self-employment, as well as employment in the for-profit, non-profit, and government sectors (our most common employment categories). We found a weak association between for-profit employment and graduation year, with recent grads more likely to work in this sector, but this was due in large part to a jump in for-profit employment among our 2015 grads (a disproportionate number of whom came from disciplines like Psychology, where for-profit employment is particularly common). We did not identify a relationship between graduation year and any other employment category.
In summary, while we are unable to say anything about causality given the structure of our dataset, we do see a persistent relationship between graduation year and higher education employment patterns. No such relationship was found in other employment sectors.
Thanks for the feedback! Feel free to send along any other questions about our data.
We recently received a question about how time is handled in our dataset, so we put our upcoming geographic analyses on the back-burner to take a quick look at how time on the job market influences employment patterns. As a reminder, we only have data for each of our grads at one point in time; while this prevents us from determining the specific employment trajectory of a given individual, we can use the full sample to examine the association between graduation year and employment type.
A simple logistic regression tells us, somewhat predictably, that the odds of faculty/professor employment are lower among more recent grads than among grads who have spent more time on the job market. Similarly, the odds of other types of higher education employment (postdoctoral fellowships, research fellowships, etc.) are higher among more recent grads. When we examine this graphically, we find that the associations are most pronounced at the upper and lower ends of our graduation year continuum; the proportion of grads from 2007 to 2010 working in either category is essentially equivalent, as illustrated in the plot below.
The relationship between graduation year and higher education employment is similar across every discipline with sufficient data, and the trends are consistent for both men and women:
We also looked at the relationship between graduation year and self-employment, as well as employment in the for-profit, non-profit, and government sectors (our most common employment categories). We found a weak association between for-profit employment and graduation year, with recent grads more likely to work in this sector, but this was due in large part to a jump in for-profit employment among our 2015 grads (a disproportionate number of whom came from disciplines like Psychology, where for-profit employment is particularly common). We did not identify a relationship between graduation year and any other employment category.
In summary, while we are unable to say anything about causality given the structure of our dataset, we do see a persistent relationship between graduation year and higher education employment patterns. No such relationship was found in other employment sectors.
Thanks for the feedback! Feel free to send along any other questions about our data.
Discussion