US Census Bureau Post-Secondary Employment Outcomes (PSEO): Doctoral Universities

In the lifecycle of colleges and universities, the month of May has historically marked the end of many students’ journey through their academic pursuits and into the beginning of their careers. While most attention to student outcomes within higher education has remained focused on retention/persistence and graduation, the topic of post-graduation employment outcomes have come to the forefront in recent years. With millions of students in America receiving diplomas this month, we turn our attention to a recently-developed data source from the US Census Bureau that provides Post-Secondary Employment Outcomes (PSEO) data for 10 partnership states, including Texas, with an additional seven states in the data pipeline.

Many states have developed systems that link higher education graduation data with workforce data to provide aggregated results of whether and to what extent graduates of public institutions are working within that particular state. The Texas CREWS website serves this purpose for Texans. Although the majority of graduates from Texas public institutions work in the state after graduation, one of the limitations to the Texas CREWS data is that any graduate who is not captured in the unemployment insurance data maintained by the Texas Workforce Commission (TWC) does not get counted. Numerous reasons for these omissions exist, but the primary one is due to graduates working outside of Texas after graduating from a Texas institution. The partnership between the Texas Higher Education Coordinating Board (THECB) and the US Census Bureau overcomes that limitation by providing earnings and wage data at a national level. For further exploration, the PSEO Explorer is a great tool for drilling into the employment outcomes data for specific institutions. However, making comparisons across institutions is currently not possible. This blog serves as an initial attempt to compare post-graduation employment outcomes across institutions within Carnegie Classifications represented in the current data.

Earnings Data: Bachelor's Degreeholders from Doctoral Institutions

The 10 partner states currently included in the PSEO data are Colorado, Connecticut, Louisiana, Maine, Michigan, New York, Ohio, Pennsylvania, Texas, and Wisconsin. States are at different levels of integration within the PSEO structure, as data are not available for all public institutions in each partner state. However, for the institutions included in the data, there is a high knowledge rate for graduates of those institutions. Data on the PSEO Explorer interactive website can be viewed for various degree levels (bachelors and masters), cohorts from 2001 through 2015, earnings percentiles, years post-graduation (1, 5, and 10), and academic programs.

For the purposes of our exploration, this post concentrates on average earnings for bachelor’s degreeholders at the 50th percentile. Institutions are clustered based on their Carnegie Classification, as the focus of this blog post is on doctoral-classified universities, and masters colleges and universities will be presented in the next post. Average earnings are for all cohorts across all instructional programs. The group of visualizations below shows each of the three Carnegie Classifications of doctoral institutions separately: Very High Research (VHR), High Research (HR), and Doctoral/Professional. The time-points measured in the PSEO data are for 1 year, 5 years, and 10 years post-graduation. The default view for each visualization has the institutions sorted from highest to lowest in the first column that represents “Year 1 Average Earnings (50th Percentile)” as shown at the bottom of the column. The second and third columns are “Year 5” and “Year 10” earnings respectively, with the fourth column representing the “percentage (%) change from year 1 to year 10” at each institution. The overall averages are weighted averages based on the number of graduates employed in Year 1 of the data. The color-coding of the bars is based on the institution’s percentage change value in the fourth column, with darker colors representing less change than lighter colors with more change over time. The visualization can be sorted using any of the four columns by hovering over the column label at the bottom and clicking on the “filter” icon that pops-up.

Doctoral Institutions: Very High Research (VHR)

For Year 1 Average Earnings, University of Texas at Arlington graduates had the highest average earnings at $48,359, with Louisiana State University graduates having the lowest average earnings at $26,475. The overall average in Year 1 was $38,003.
For Year 5 Average Earnings, University of Texas at Austin graduates had the highest average earnings at $61,338, with Colorado State University graduates having the lowest average earnings at $46,799. The overall average in Year 5 was $54,632.
For Year 10 Average Earnings, University of Michigan graduates had the highest average earnings at $81,800, with University of North Texas graduates having the lowest average earnings at $58,374. The overall average in Year 10 was $69,978.
For the percentage change from Year 1 to Year 10, Louisiana State University graduates increased by almost 142%, with the University of Texas at El Paso having the smallest percentage increase of 38.7%. The overall average percentage change was 85.6%.

Doctoral Institutions: High Research (HR)

For Year 1 Average Earnings, Colorado School of Mines graduates had the highest average earnings at $60,838, with Louisiana Tech University graduates having the lowest average earnings at $26,828. The overall average in Year 1 was $34,416.
For Year 5 Average Earnings, Colorado School of Mines graduates had the highest average earnings at $81,682, with University of Louisiana at Lafayette graduates having the lowest average earnings at $42,844. The overall average in Year 5 was $48,158.
For Year 10 Average Earnings, Colorado School of Mines graduates had the highest average earnings at $99,860, with University of Louisiana at Lafayette graduates having the lowest average earnings at $51,772. The overall average in Year 10 was $58,516.
For the percentage change from Year 1 to Year 10, Louisiana Tech University graduates increased by over 113%, with the University of Texas – Rio Grande Valley having the smallest percentage increase of 31%. The overall average percentage change was 71.4%.

Doctoral Institutions: Doctoral/Professional

For Year 1 Average Earnings, Texas Woman’s University graduates had the highest average earnings at $42,855, with University of Louisiana at Monroe graduates having the lowest average earnings at $28,071. The overall average in Year 1 was $35,374.
For Year 5 Average Earnings, Texas Woman’s University graduates had the highest average earnings at $53,786, with University of Northern Colorado graduates having the lowest average earnings at $42,194. The overall average in Year 5 was $47,479.
For Year 10 Average Earnings, Texas Woman’s University graduates had the highest average earnings at $61,176, with Texas A&M University-Commerce graduates having the lowest average earnings at $49,895. The overall average in Year 10 was $55,018.
For the percentage change from Year 1 to Year 10, University of Louisiana at Monroe graduates increased by over 85%, with the University of Texas at Tyler having the smallest percentage increase of 32.6%. The overall average percentage change was 57.7%.

NOTES: If you are unable to read all of the labels in the visualization below, we recommend clicking the “Full Screen” button on the lower right of the visualization window to enlarge the information. Hovering over each bar will provide a tooltip with additional information.

Comparison between Earnings and In-State Employment Rates

Another data point included in the PSEO data is the count of graduates who were employed within the state where they earned their bachelor’s degree. By dividing this in-state employment count by the total number of earners, we can determine an in-state employment rate. In the PSEO Technical Documentation, the researchers indicate that “mobility and higher wages are positively correlated” (p. 10), which would mean that the higher the percentage of in-state employment for a university, the lower the average earnings would be for their graduates. To examine whether this statement was supported by the data on doctoral-level institutions, we plotted each of the doctoral universities in the PSEO dataset using scatterplots with “Average Earnings” on the x-axis and “Percentage Employed In-State” on the y-axis, as shown in the visualizations below. The institutions are color-coded based on their state, and each time-point (Year 1, Year 5, and Year 10) are plotted in separate panels. The trend line through the data points is considered to be the “line of best fit.”

Doctoral Institutions: Overall Trends

In general, it appears that there is a positive correlation between higher mobility (lower in-state employment rates) and earnings. We can see that the trend line in each panel slopes downward from left-to-right, meaning that as the percentage of graduates who work out-of-state increases, the average earnings also increases.
HOWEVER, notice the blue dot to the far right on each of the panels: Colorado School of Mines. In these doctoral-level PSEO data, this institution would be considered an outlier that is skewing the results downward. While not shown below, if we remove Colorado School of Mines from the data, the Year 1 slope actually reverses to a positive value, meaning that higher in-state employment rates lead to higher salaries. In Year 5 and Year 10, the downward (negative) slope returns and follows a similar pattern as shown below, which would support the PSEO statement related to higher mobility and earnings.
As the length of time from graduation increases, so does the steepness of the slope of the trend line. Notice that the Year 1 line is relatively flat from left-to-right, while the Year 10 line is dramatically steeper. This would seem to indicate that the correlation between mobility and earnings increases over time to the point of being statistically significant.
For Texas public universities, all institutions are either at or above the line of best fit for the Percentage Employed In-State within each time panel. Sam Houston State University has the highest percentage of in-state employed graduates in each time panel (93.1% in Year 1, 91.9% in Year 5, and 90.2% in Year 10). On average across Texas doctoral public institutions, 89% of bachelor’s degreeholders stay in Texas to work during Year 1 after graduation, with that average falling to 85.7% in Year 5 and 83.8% in Year 10.

Doctoral Institutions by Carnegie Classification

When factoring Carnegie Classification into the equation, we see slight variations in trend lines compared to the overall scatterplots, which means there is an interaction effect taking place in these data.
For the Very High Research Activity institutions (top row), the slope is positive for Year 1 and Year 5, and drops to negative in Year 10. It is not until the University of Michigan graduates (lowest orange dot) drop below 30% in-state employment does the mobility and earnings correlation statement hold true.
For High Research Activity institutions (middle row), the inclusion of Colorado School of Mines continues to skew the data downward. When removing this institution, the trend line is positive in Year 1, relatively flat in Year 5, and negative in Year 10 for institutions in this middle category.
For the Doctoral/Professional institutions (bottom row), the Texas trends shown in the overall data have the direct result of pulling the trendlines in the positive linear direction across all three time points.

NOTES: If you are unable to read all of the labels in the visualization below, we recommend clicking the “Full Screen” button on the lower right of the visualization window to enlarge the information. Hovering over each point will provide a tooltip with additional information.

So What?

The most important benefit of these PSEO data is the ability to make comparisons between and among institutions across state lines using a common metric that is not limited by within state borders data systems. The PSEO data also provide the opportunity to see these trends over time. Using the “Scatterplots: Doctoral by Carnegie” visualization above, we can see how the most prominent public institutions, generally known as “R1” institutions, in each of these 10 states perform against one another in terms of the earnings of their bachelor’s degree graduates. What we can also see is that, excluding the Colorado School of Mines, graduates from institutions in the Very High Research Activity category tend to earn more at each time point of measurement, although the distribution of mobility (percentage employed in-state) is relatively similar over time and across institution types.

This blog post is just the proverbial “tip of the iceberg” in terms of what can be done with these PSEO data. We strongly encourage you to visit the PSEO Explorer website to further explore institution-specific data that allow for drilling-down into earnings data at the academic program level, as well as a very nice feature showing how graduates “flow” from academic program preparation to specific industries. This feature alone would be greatly beneficial for expanding our knowledge about what paths our graduates follow after earning their degrees at our institutions.

If you are wondering about the data for Master’s Public Colleges and Universities, stay tuned. The May 15th blog will present those data in similar form as has been presented here for doctoral-level institutions.