A plot of the National Institutes of Health’s National Research Service Awards Year 0 stipend by Financial Year. Also includes a comparison of salaries with their approximate value in 2017, using the Personal Consumer Expenditure Index.
In 2016, the very earliest days of Future of Research’s existence as a nonprofit were dominated by the announcement of updates to the Fair Labor Standards Act, and in particular how that would effectively raise postdoc salaries to $47,476 on December 1st 2016.
The birth – and death – of this update to the Fair Labor Standards Act, and how it was being implemented at institutions, occupied much of our attention, and is summarized in our publication Monitoring the compliance of the academic enterprise with the Fair Labor Standards Act. But even though the update was ultimately not implemented, the academic research system largely went ahead with changes to institutional policies to raise recommended postdoc salaries.
We were however aware of the issue that institutions vary significantly in their ability to count, and presumably, identify postdocs. This led us to ask a number of questions:
- If institutions are unable to count their postdocs, and presumably are not overseeing them, do all postdocs receive the salaries set out in an institution’s policy?
- How strong is the relationship between the National Institutes of Health National Research Service Award stipends (which affect only 15% of graduate students and postdocs funded by NIH, which is not the only funder of postdocs) and what postdocs are getting paid?
- Are there any factors affecting salary, such as location, gender, or job title?
We therefore began gathering data on postdoc salaries, which we presented in a raw form in our Investigating Postdoc Salaries Resource, and have now analyzed and published as a peer-reviewed publication: Athanasiadou et al. 2018, “Assessing the landscape of US postdoctoral salaries” in Studies in Graduate and Postdoctoral Education*.
What did we do?
To get postdoc salaries, we used a rather blunt instrument: Freedom of Information requests at public universities, asking for the job title and salary on December 1st 2016 of each postdoc.
We used this methodology, rather than for example asking individual postdoctoral officers or administrators for data, for a variety of reasons:
- It was simply the fastest way of requesting data, and we had limited resources to commit to the effort;
- It provides a standard methodology – and therefore reduces variation due to things like personal relationships with individual staff members;
- It provides an interesting internal measure of whether an institutional administration can communicate internally to provide relatively trivial data about postdocs.
While requesting only job title and salary, a large amount of the data was also supplied with names and departments. We therefore analyzed the salary data looking at geographic location; discipline/field of study; gender (using an algorithm to attempt to assign gender to names); job title; and the amount of funding received by the institution from the National Institutes of Health and the National Science Foundation.
Why did we do it?
In addition to the reasons laid out above, we carried out this analysis to first determine how postdoc salaries looked generally, and to see whether there were further questions we could ask in the future with longitudinal and more detailed studies.
We are also interested in which lever to pull to affect policy changes. Raising postdoc salaries is a recommendation consistently made to improve both the lot of the postdoc, but also to divorce cheap labor from training in academia. It is a very simple intervention that could be easily measured, and in particular we wanted to see whether there were particular stakeholders or agencies to target to ensure that recommended salary raises occur.
Our interest in postdoc salaries is firmly based on the observation that people are leaving academia on the basis of financial factors – for example, being able to afford childcare – and we strongly believe that science cannot be a meritocracy while selection factors such as ability to withstand financial hardship (among others) are in play.
Especially when such factors are vigorously defended by those who stand to benefit from a pool of cheap labor, rather than prioritizing selection of scientists based on talent, we see it as an important part of our work to talk expressly about money. We see this project as a step towards greater financial transparency in academia.
What did we see?
NIH policies determine US postdoc salaries
The most striking result was that the national *median* postdoc salary is determined by the *minimum* (Year 0) NIH NRSA postdoctoral stipend – over 20% of all salaries were in a $25 range around the NIH’s NRSA Year 0 stipend at the time. This is not surprising – most institutions simply benchmark their salary policies for all postdocs to the NIH NRSA policy. It points to a policy lever to pull – NIH is key to raising all postdoc salaries in the US.
Changes to NIH NRSA stipend over time. Key events and policy recommendations, and when NIH implemented these changes – if at all – are identified, both in $ amounts identical to the year of recommendation, and when they were achieved when adjusted for inflation (e.g. “1994 Recommendation” box shows that the proposed $25k stipend was reached in 1999, but accounting for inflation it was not truly implemented until 2001.)
Postdoc salaries exist over a wide range
We saw salaries across a wide range. One factor was that some institutions reported data that was not an annualized salary, but rather an artefact of some institutional policy e.g. a fellowship paid directly to the postdoc may have appeared as a salary of zero. However, other institutions were able to provide the data we requested, and so we had to take any salary above $23,660 (the legal minimum) at face value, while presuming that those below $23,660 were errors and not a violation of federal labor law. That we found salaries at exactly $23,660 suggests that it was correct to use this number as the lowest threshold.
There were relatively few salaries in the lower range; we were interested to see that there were also a large number of salaries well above the NIH level. We did not find a field dependence to this – it is not correct that humanities postdocs have lower salaries, for example – and anecdotally we suspect this may be due to a significant number of postdocs negotiating salaries (the plot below by Nature came out when we originally released our raw data), in addition to our finding that salaries are higher in the West Census Region.
There is a gender pay discrepancy in postdoc salaries
We found a gender pay discrepancy based on gender nationwide. When breaking down by Census region, this gender pay discrepancy is concentrated in the NorthEast and South; there is no significant discrepancy in the MidWest, or the West.
Postdoc salaries by gender and by U.S. Census Region.
Salaries vary with job title descriptor
Various terms used in the myriad of titles for postdocs were associated with an adjustment in the value of a postdoc’s salary:
Advocating for change
We have actively used and encourage use of this data on national initiatives, such as the Next Generation Researchers Initiative study at NASEM, to advocate for raises to postdoc salaries.
Empowering postdocs to negotiate and share salaries
We have worked together with Personal Finance for PhDs to produce postdocsalaries.com – where postdocs can self-report salaries, and benefits, and prospective postdocs can compare salary information for potential institutions. This also contains the ability to comment on experiences regarding salaries. Currently about 25% report negotiating their salary.
It is very clear that postdocs are negotiating and making decisions based upon salaries. We continue to challenge the notion that your financial situation should be irrelevant in pursuing a research career, and instead a true meritocracy would ensure that finances are not a barrier to retaining the best talent.
Continuing the research
We are currently repeating and extending our data collection effort for salaries on December 1st 2017 and 2018:
- In 2016, we asked institutions with >300 postdocs in science and engineering (based on NSF data) for data, and this on occasion included other institutions within the same system with smaller populations. We have now extended this to all institutions with >75 science and engineering postdocs, and ensuring that each state is represented (except Pennsylvania**).
- We are asking explicitly for name and department, and also asking if gender information can be disclosed.
If you have any questions about the paper, please feel free to comment below, or on Twitter, or email firstname.lastname@example.org
*Please note that we are currently addressing issues with the publication, including incorrect authorship order and use of a previous version of Figure 1, introduced by the journal. Errors do not affect the data or findings of the paper.
**Pennsylvania currently exempts public institutions (University of Pittsburgh, Penn State, Temple and Lincoln Universities) from meeting Freedom of Information requirements.