Introduction

Significant financial resources are invested towards fellowships for doctoral students in STEM fields. For example, the FY24 requested budget for U.S. National Institutes of Health predoctoral and postdoctoral research training grants was over $1 billion [1], while the request for the U.S. National Science Foundation Graduate Research Fellowship Program totaled $380 million [2]. The return on investment is reaped both through the direct research output of current doctoral students as well as the future workforce contributions of PhD earners who augment research capacity through roles in academia, government, and private industry. In addition, doctoral students who have teaching responsibilities contribute to the education system by increasing an institution’s instructional capacity. Across industries, PhD holders represent a highly skilled and specialized segment of the workforce that augments economic growth.

Despite such a large investment in doctoral education within engineering, both in terms of financial and human resources, surprisingly little research has explored this phenomenon in depth within engineering and its highly varied sub-disciplines. For example, the Council of Graduate Schools’ (CGS) annual report on graduate student application, enrollment, and degree conferrals includes eleven broad fields of study [3]. Similarly, the annual report on doctoral education created by the National Science Foundation includes nine fields of study across academic fields [4]. Research at the doctoral level [5] showed differences between engineering, physical sciences, and life sciences in terms of students’ funding mechanisms. The researchers highlighted the existence of variation within the broad engineering field but did not explore sub-disciplines to explain that variation. Key characteristics of biomedical/biological engineering, such as research funding sources and the importance of postdoctoral training, are more similar to life sciences than to the engineering sub-disciplines within which biomedical engineering is so often combined for policy analyses and organization within higher education institutions.

Using the National Science Foundation’s (NSF’s) Survey of Earned Doctorates (SED) dataset, our research moves beyond such field-level analyses to interrogate how doctoral student funding mechanisms vary across engineering sub-disciplines, with a particular emphasis showing distinct differences for biomedical and biological engineering relative to the other sub-disciplines. Graduate student funding mechanisms have been shown to relate to important outcome variables including degree completion, time to degree, learning outcomes, and short- and long-term career outcomes [6,7,8,9,10,11,12,13,14]. Given these important connections, our analysis of within-engineering variation for funding would be of interest to many stakeholders, including, for example, engineering college-level graduate education administrators seeking to develop a strategy for sustaining equitable funding models, department-level graduate program directors seeking to understand contexts that might be most likely to have transferable ideas, or student support programs across a variety of engineering departments.

Our study focuses on three of the most common funding mechanisms for doctoral students in engineering: fellowships, teaching assistantships, and research assistantships. Given prior research, we anticipated that differences in engineering sub-disciplinary organization and culture can result in differing availability and emphases of different funding mechanisms, which has several implications for graduate students’ experiences and eventual career outcomes. Our first research question is as follows:

RQ 1

How do funding mechanisms vary across engineering sub-disciplines?

Our second research question considers how the engineering sub-disciplines fit within the broader STEM context. For example, all engineering sub-disciplines could differ from disciplines in other STEM fields and behave more like the parent engineering field in aggregate. Alternatively, the within-field variation could mean that certain engineering sub-disciplines are more similar to some of the other STEM disciplines than other engineering sub-disciplines. As described by Biglan [15, 16], engineering as a field is an applied form of pure science fields, and so it is possible that sub-disciplines like chemical engineering might be more similar to the pure field of chemistry than to other engineering sub-disciplines, which could have implications for how graduate experiences could be structured within and across colleges. With this difference in mind, we explore this potential with our second research question:

RQ 2

How does variation in funding mechanisms across engineering sub-disciplines map onto the larger STEM disciplinary landscape?

Review of Prior Literature

We draw on broader research on graduate education to discuss field-level differences and why those differences matter, in particular with respect to funding mechanisms. Much of the theoretical basis for the examination of disciplinary differences in doctoral education began with the work of Biglan [15, 16] and Becher [17], who sought to classify the cultural and social structures of academic disciplines. The resulting frameworks identify underlying cultural identities and paradigmatic assumptions that vary across disciplinary boundaries that can manifest in critical differences in doctoral practices. Disciplines tend to be the most important drivers of faculty members’ attitudes and behaviors [18, 19] and tend to be much more important determinants of cultures and processes than institutions [20]. In other words, for example, a biomedical engineering department at Institution A and a biomedical engineering department at Institution B will likely be more similar to one another than the biomedical engineering department and English department at Institution A. It is very difficult for any one institution to challenge this kind of organization because the academic disciplines enable faculty and graduate student mobility across institutions [21].

Because of the importance of disciplines, and in contrast to many studies of undergraduate students in which experiences are often examined at the institutional level, the department is often the central focus of the doctoral student experience [22,23,24,25]. A key feature of doctoral education is focused on socializing students to disciplinary norms [23, 24]. Departmental practices regarding doctoral education tend to be shaped by the disciplinary norms of faculty members [26]. Academic programs within departments tend to determine most policies that affect doctoral student life, such as admissions, curriculum, degree requirements, and financial support [27]. This financial support component is the key focus of our study because of the many different implications that the kind of funding has for students’ experiences and outcomes.

As previously demonstrated [5], doctoral student funding portfolios varied substantially across STEM disciplines. That prior work showed differences in the extent to which doctoral students were supported financially via different mechanisms, including research assistantships, teaching assistantships, and fellowships, for example. The authors posited that the nature of those fields as well as resource availability and pressures helped explain why funding portfolios varied across the disciplines. Much of a STEM doctoral student’s experience in their program can be driven by the ways in which they are funded, the implications of which are extremely important to understand and be mindful of in the design of programs.

Students funded via research assistantships gain access to opportunities to develop research skills [10, 28] and have greater access to research labs [10, 29, 30], which enable opportunities for increasing scholarly metrics during graduate school and aligning dissertation work with assistantship work [10, 13, 30, 31]. Although some researchers have highlighted situations in which having a research assistantship does not guarantee access to all of the potential benefits of this kind of experience [28, 30, 32, 33], this form of funding has been linked to increased research productivity, agency, and prestige [6, 9, 10, 29].

In contrast, students funded via teaching assistantships have opportunities to build their socialization within the discipline by interacting with faculty and peers [10, 34]. Being able to develop teaching expertise via this kind of appointment can help position doctoral students well for future faculty careers [35,36,37], particularly when the funding source is interspersed for a semester with a different, longer-term appointment type like a research assistantship or fellowship [6]. Research has shown that students funded via a teaching assistantship have opportunities to develop a better understanding of their discipline and the ways in which curricula and research are designed, build self-confidence and organizational skills, and also experience enhanced communication skills [10, 32, 38,39,40]. However, in contrast to research assistantships, dissertation work tends to not be aligned with teaching assignments, and thus there has been evidence of increased time to degree for students funded on this assistantship type [6, 10, 13, 41] as well as increased likelihood of not having a job offer upon graduation for these doctoral students [42].

Finally, students who are funded via fellowships tend to have their financial requirements covered without an associated work expectation [26, 34, 43], and there is a perceived level of prestige tied to this funding type [33]. Although these students typically have greater autonomy and time dedicated to their own dissertation work [44, 45], there is some evidence that students can have a harder time gaining access to research opportunities and faculty support [26, 34, 43]. Recent research [10] showed lesser development of research skills, teamwork/project management skills, peer mentoring and training skills, and communication for students funded via external fellowships relative to students funded via internal fellowships and research assistantships. Biomedical sciences PhDs funded through fellowships were less likely to select research-focused jobs for their post-graduation employment than PhDs funded through RAs [46]. It is also unclear how fellowships impact degree progress. One study [29] found that doctoral students funded on fellowships had increased odds of reaching candidacy but decreased odds of completing their degrees. Another study [34] found positive relationships with retention for PhD students funded on fellowships, although the sample may have been biased by competitive fellowship selection criteria. Others [42] showed potentially troubling job attainment patterns for Black/African American STEM doctoral earners who were funded via fellowships, which is concerning since historically excluded groups tend to be overrepresented among fellowship recipients [13].

In sum, this prior research has demonstrated how academic fields and disciplines vary in terms of their norms and practices pertaining to doctoral education, which includes doctoral student funding mechanisms. In addition, a growing body of evidence has shown the importance of funding mechanism for students’ experiences and subsequent outcomes. Although a few prior studies have connected these two literature spaces (i.e., disciplinary differences and doctoral funding mechanisms), the prior work has remained at the field level (i.e., engineering as a whole or STEM) as opposed to interrogating within-engineering differences across sub-disciplines. Our work makes an advance by looking within engineering.

Prior attempts to examine within-engineering variation are most common at the undergraduate level. Foundational qualitative work on engineering disciplinary subcultures [47] spurred follow-up studies that examined the characteristics of students and faculty and student persistence and completion in specific engineering sub-disciplines including chemical engineering [48], civil engineering [49], environmental engineering [50], and mechanical engineering [51]. Other comprehensive studies have also sought to compare engineering sub-disciplines along similar variables [48, 51] as well as curricular and instructional environments [52]. Like this research at the undergraduate engineering level, we similarly would expect sub-disciplinary differences across students’ experiences and outcomes at the doctoral level.

Research exploring differences between disciplines at the doctoral level in engineering is fairly rare. Artiles et al. [53] demonstrated differences in doctoral advisor selection processes between different sub-disciplines within engineering, science, and math that could largely be explained by programs’ availability of different funding mechanisms. Although biomedical, biological, and biosystems engineering programs were not in that prior work, some of the findings related to chemical engineering may be relevant given some of the similarities between the sub-disciplines. Chemical engineering programs in the sample all funded students centrally via fellowships for at least one semester, which enabled students to complete engagements across the program with a variety of potential faculty members followed by a “matching” approach to advisor selection. For other engineering sub-disciplines, students tended to have specific advisors prior to matriculation because funding tended to be more decentralized and controlled by individual faculty members. Additional work by Artiles and Matusovich [54] focused on this unique advisor selection process within chemical engineering to understand tradeoffs associated with this approach in greater detail. Unfortunately, this work does not evaluate outcomes of advisor selection processes nor student success. These results further motivate the need to understand doctoral education better at the sub-discipline level. Most relevant to the readership of this journal, biomedical and biological engineering was not included in those prior study’s samples. We anchor our analyses on this sub-discipline in our comparisons of how funding mechanisms vary across the engineering sub-disciplines.

Data and Methods

We used data from the NSF’s Survey of Earned Doctorates data set managed by the National Center for Science and Engineering Statistics (details available at https://ncses.nsf.gov/surveys/earned-doctorates/2022). The SED is characterized by comprehensive coverage of doctoral recipients from institutions in the United States and has collected information from research doctoral recipients (PhD) from accredited institutions within the United States continuously since the 1957–1958 academic year. Using a combination of self-administered paper surveys, web-based surveys, and computer-assisted telephone interviews, graduate schools typically collect SED responses at the time of degree completion. For the 2012 SED (i.e., the midpoint of our period of record), 92% of the 51,008 recipients of doctorates completed the survey; non-respondents were reconstructed from any other available information. The period of record to which we had access for our analysis included responses from individuals from fiscal years 2007 to 2016.

We filtered the full sample first to respondents who earned a doctorate in engineering or computer and information sciences (N = 103,373). Next, we derived an engineering sub-discipline variable based on a self-reported 3-digit code that “best describes the primary field of dissertation research.” Some of these codes were combined into similar sub-disciplines to reduce low counts in certain low count sub-disciplines; Table 1 displays frequencies of the sample by engineering sub-discipline.

Table 1 Sample of respondents, by engineering sub-discipline (shorthand in parentheses)
Full size table

We explored three funding variables, all of which were derived from a single-choice survey item on the SED asking respondents to indicate their primary type of funding mechanism that supported their graduate education. The single item contained fourteen possible choices, including Fellowships and Grants (which we combine, consistent with [5], since these mechanisms function in similar ways in which they are awarded to individual students; these are not to be conflated with research grants awarded to institutions and managed by faculty), Teaching Assistantships, and Research Assistantships (Table 2). This study focuses on these funding mechanisms because they dominate engineering doctoral students’ funding portfolios [5].

Table 2 Summary counts of students’ primary funding mechanism
Full size table

We conducted an analysis of variance with Bonferroni post hoc comparisons to address the first research question, which examined variation in funding across the 11 largest sub-disciplines. Although we present overall percentages for each sub-discipline in the Results section, an analysis of variance enabled us to account for any funding portfolio differences across institutions for each sub-discipline (i.e., comparing the national set of individual biomedical/biological engineering programs with the national set of individual mechanical engineering programs, etc.). Separate analyses were run for each of three funding variables since those were mutually exclusive options for respondents. In presenting our results, we use biomedical/biological engineering as the reference comparison discipline. To sort 112 STEM sub-disciplines by funding variables and address the second research question, we conducted a k-means cluster analysis with STEM sub-discipline as the unit of analysis. The clustering was conducted on percentage variables with that particular funding mechanism. Clustering was conducted on five centers, as determined through a scree test following the elbow curve method; each data point was assigned to a cluster to which the distance was nearest using simple Euclidean distance since each variable was on the same scale.

Limitations

In interpreting findings, there are some key limitations to consider. First, the SED is only completed by individuals who complete their doctorates. It does not include information about Master’s students, currently enrolled doctoral students, or students who depart their institution, and so these results should be interpreted as funding portfolios of successful doctorate earners. Second, we focused our analysis on students’ primary funding mechanism as opposed to operationalizing the comprehensive set of ways that individual students were funded, which is consistent with prior studies of students’ funding [5]. Because the SED is a national-scale data set whereby information is reported by students from all institutions, it is unable to collect precise data about the timing, sequencing, or length of each kind of funding; therefore, our analysis focused on the “primary” funding variable. Completing an analysis of all sources of funding during a student’s time as a graduate student would require institution-level data to be able to make claims with any degree of confidence. For example, a student’s secondary source of funding could mean a $2000 internal fellowship as part of a recruitment mechanism or a single year of internal fellowship support, which would function very differently, but is indistinguishable in the SED. The primary funding source, therefore, is the only funding variable in the SED with sufficient information to make claims about the research questions in a confident manner. Other work has explored a more precise analysis at the single institution level, which is beyond the scope of our study, but even that prior single institution work found challenges in seeking greater levels of precision [55]. This is an important direction to consider in future work. Third, survey respondents self-report on all items of the SED, including their funding, and so there is always the opportunity for students to interpret questions differently or not understand their actual funding mechanism. Finally, we also decided to smooth out potential year-to-year variability by focusing on a five-year period of record, but it is possible that this period of record may not reflect any changes in funding opportunities that occurred following COVID-19.

Results

Doctoral Student Funding Mechanisms Across the Engineering Sub-disciplines

Results shown in Figure 1 and Table 3 address the first research question. For biomedical/biological engineering as a whole, 42% of doctoral recipients were primarily funded via fellowships, 46% were primarily funded via research assistantships, and 4% were funded via teaching assistantships. The remaining 8% of doctoral recipients in biomedical/biological engineering had their primary funding come from the other sources noted in Table 2. Figure 1 and Table 3 demonstrate how the funding portfolios vary quite substantially across the engineering sub-disciplines—these kinds of differences are not evident in analyses that aggregate engineering as a singular field.

Fig. 1
figure 1

Proportion of doctorate earners funded via the three largest primary funding mechanisms across engineering sub-disciplines

Full size image
Table 3 Funding portfolios of primary funding mechanism1 across engineering sub-disciplines
Full size table

Compared to every other engineering sub-discipline, using an ANOVA with post hoc comparisons, a greater percentage (p < .05) of biomedical/biological engineering doctoral students were funded via a fellowship (42%). Consequently, biomedical/biological engineering had significantly lower proportions of students supported via research (46%) and teaching assistantships (4%) than nearly all other fields, the exceptions being for proportions of students funded via teaching assistantships in chemical engineering (5%) and materials science engineering (5%). Environmental engineering (30%), chemical engineering (27%), and materials science engineering (24%) also had high proportions of students funded via fellowships compared to other disciplines, yet lower than the proportion for biomedical/biological engineering (42%). Conversely, civil engineering (17%), electrical engineering (14%), and industrial engineering (13%) had the lowest percentages of students funded via fellowships.

When examining disciplinary differences in proportion of students funded via teaching assistantships, industrial engineering (18%) had the highest percentage funded this way. Computer science (14%) also had higher percentages of students funded via teaching assistantships compared with all disciplines except engineering mechanics (17%) and industrial engineering (18%). Unlike for fellowships and teaching assistantships, there was not a clear disciplinary outlier for research assistantships. Electrical engineering (67%) was the highest among all of the disciplines, but it was clear that research assistantship funding is the predominant mechanism for doctoral earners across all engineering sub-disciplines (ranging from 43 to 67%).

Characterizing Engineering Sub-discipline Funding Portfolios Within STEM Broadly

Using a k-means cluster analysis, we next examined how STEM disciplines group together according to funding variables in addressing the second research question. We named each cluster to characterize average funding portfolios across the sub-disciplines comprising each cluster relative to the average funding portfolios of sub-disciplines comprising the other clusters (unweighted by number of students). The cluster analysis indicated five groupings, three of which had one funding type that was more prevalent than other types: (1) High Fellowships, (2) High GRAs/Low GTAs, and (3) High GRAs/Balanced GTAs & Fellowships. The remaining two groups maintained more balanced funding portfolios, differing in which funding type was lowest: (4) Low GRAs/Balanced GTAs & Fellowships, and (5) Low GTAs/Balanced GRAs & Fellowships. We first present and analyze our results by focusing only on engineering sub-disciplines (Table 4), and then discuss how they clustered with other STEM disciplines (Table 5).

Table 4 Clustering on funding variables for engineering sub-disciplines only
Full size table
Table 5 Clustering on funding variables for STEM sub-disciplines
Full size table

The majority of engineering sub-disciplines (22 out of 31; 71%) grouped together into a cluster with High GRAs/Low GTAs. In this way, most engineering sub-disciplines seem to emphasize research assistantship funding models compared with teaching assistantships, which account for only eight percent, and fellowships, which account for 19% of funding mechanisms in this cluster. In fact, the proportion of fellowships in this cluster (19%) is the lowest proportion across all five clusters and matches the lowest proportion of teaching assistantships with High Fellowships cluster. Conversely, the cluster with the highest proportion of teaching assistantships (31%)—Low GRAs/Balanced GTAs & Fellowships—did not have any engineering sub-disciplines at all, further illustrating the generalized emphasis of research assistantships over teaching assistantships across most engineering disciplines. Exceptions are industrial engineering, operations research, and computer and information sciences, which all clustered in the High GRAs/Balanced GTAs & Fellowships, with the second highest percentage of teaching assistantships (23%) and research assistantships (44%). Biomedical/biological engineering was alone among engineering disciplines to sort into the High Fellowships cluster, indicating considerably more prioritization of fellowships (47%) than all other clusters, which as we describe later, seems to relate to disciplinary similarities with biological sciences. The final cluster—Low GTAs/Balanced GRAs & Fellowships—which included disciplines with a balanced funding portfolio that deemphasized teaching assistantships (10%), included more generalized engineering or computer information concentrations (i.e., general engineering; engineering, other; & computer information systems, other) along with systems-focused disciplinary concentrations (i.e., systems engineering & computer information systems).

Table 5 includes the same clusters as Table 4 but also displays the other STEM sub-disciplines to examine contextual similarities in funding mechanisms with disciplines outside of engineering. As noted above, biomedical/biological engineering was the only engineering discipline in the High Fellowships cluster; the majority of the biological sciences subfields are also found in this cluster, which prioritizes the highest percentage of fellowships (47%) across all of the clusters. The second cluster—High GRAs/Low GTAs—which includes the majority of engineering disciplines, also includes many physics-focused disciplines. Like engineering, physics-centric disciplines appear to utilize research assistantships more frequently than fellowships or teaching assistantships. Chemistry and related disciplines dominate the High GRAs/Balanced GTAs & Fellowships cluster joined only by biotechnology and structural biology as well as industrial engineering, operations research, and generalized computer and information systems degrees. The cluster absent any of the engineering degrees—Low GRAs/Balanced GTAs & Fellowships—includes the fewest number of disciplines (8) that all mostly engage with earth sciences. Like engineering disciplines, the final cluster—Low GTAs/Balanced GRAs & Fellowships—included a collection of generalized or other chemistry, physics, and earth science concentrations.

Discussion and Implications

We explored the ways in which primary funding mechanism for doctoral students varies across the engineering sub-disciplines. Biomedical/biological engineering behaved much more like its biological and life science peers in utilizing fellowships to fund graduate students more extensively than other engineering sub-disciplines. Fellowships are more common in life sciences [5, 34]; our study provides further evidence of this relationship in how it stretches into this sub-discipline of engineering. The majority of engineering sub-disciplines relied more on research assistantships to fund graduate study, similar to prior work that explored engineering in aggregate [5]. Our unique finding for biomedical/biological engineering points to a sub-discipline that perhaps should not be considered in aggregate with other engineering sub-disciplines when characterizing graduate education.

One potential explanation for the difference in graduate student funding portfolios for biomedical/biological engineering and other engineering disciplines is gender diversity. Biomedical/biological engineering and biological and life sciences have high representation of women, as well as a high proportion of students funded by fellowships. Thus, it is possible that these fields are skewing results that women and students of color are overrepresented among fellowship recipients. Yet, these trends have been documented across academic fields, including humanities and social sciences [13], and it is common across STEM disciplines to offer fellowship funds as a strategy for diversifying their student population; special funds are often made available to all disciplines by the institution's graduate school [56].

Another potential explanation for funding differences between biomedical and other fields of engineering is access to additional funding sources. U.S. students across engineering and STEM disciplines are funded by National Science Foundation Graduate Research Fellowship Program, Ford Foundation, and the GEM Fellowship Program. Biomedical engineering students are additionally eligible for fellowships from health-related organizations such as National institutes of Health, PhRMA Foundation, Alzheimer's Association, and American Federation for Aging Research. Private foundations with human health as a mission (e.g., Wellcome Trust, Howard Hughes Medical Institute, W.M. Keck Foundation, and Bill and Melinda Gates Foundation) may also contribute to these differences [57]. Organizations funding fellowships rarely publicize numbers of recipients or awarded funds, so the data are difficult to come by. One potential avenue for more detailed analysis is the Survey of Graduate Students and Postdoctorates in Science and Engineering managed by National Center for Science and Engineering Statistics, a census survey in which institutions report more detailed funding information on their graduate students.

As we reviewed previously, there are a variety of tradeoffs associated with each funding mechanism in terms of students’ access to experiences [9], agency in decision making [6], time to degree [8, 14], graduation and career outcomes [8, 13], and career productivity [7, 13]. As compared to those funded by research assistantships, the greatest risk to fellowship-funded students is differential access to research groups, which may delay degree progress, limit access to mentors, and lead to attrition [34, 43]. With a lower percentage of students funded via research assistantships, it is important to ensure students are able to develop research skills and connect to research labs, some of the benefits of being funded via a research assistantship [28,29,30]. Given the prominence of fellowship funding relative to other engineering sub-disciplines, we do wonder whether the same level of prestige is associated with this kind of funding mechanism within biomedical/biological engineering as has been observed in other disciplines within graduate education [33]. The literature would anticipate that these fellowship recipients would have greater autonomy and time for their own dissertation work [44, 45], but prior research would also caution that it could be harder for students to gain access to research opportunities and faculty support [26, 34, 43] as well as find opportunities to develop a range of skills because of their funding [11]. While some advisors treat students on fellowships no differently from those funded by research grants, these national studies suggest that many fellowship students experience less engagement with research at some point during their graduate study [10, 11]. We suspect that in biomedical/biological engineering, many fellowship-funded students conduct their research independent of a specific research grant with detailed goals and deadlines, but not independent of a lab and PI. Fellowship students may, for example, explore higher-risk research areas to obtain preliminary data for future proposals, but with potentially less guidance than students on more structured projects. Our results cannot speak to whether this is the case, but the distinct difference in funding portfolio that our research highlights points to an area of which researchers and program leaders and coordinators should be mindful.

Having such a large proportion of students funded predominantly via fellowships as the primary mechanism also allows affordances that other engineering sub-disciplines do not have. As prior research showed [53, 54], affordances provided by chemical engineering’s high availability of fellowships, particularly in students’ first year, had important implications for advisor selection processes. The increased flexibility allowed a prolonged matching process whereby advisors and students had opportunities to spend more time making more intentional, selective decisions about advising relationships. Biomedical/biological engineering was not in that prior work’s sample, but we would imagine that similarities would be expected for this sub-discipline. When looking for other ideas to enhance processes within graduate education inside the engineering context, biomedical/biological engineering and chemical engineering might look to one another for ideas. In considering the results of the cluster analysis, a range of life science disciplines actually look more similar to biomedical/biological engineering than the other engineering sub-disciplines from a doctoral student funding portfolio perspective. These STEM sub-disciplines could be the best potential partners for ideas around enhancing graduate education.

These results point to several directions for future research. The Survey of Earned Doctorates does not include many questions about funding, so institution-level data sets are needed to explore questions about different funding types and sequencing those types to inform graduate education leaders making decisions about fellowship and other funding awards. Similarly, most data sets do not distinguish between internal and external fellowships, yet research has uncovered substantive differences in how these impact students [10, 58]. Prior studies hint at the relationship between graduate student funding and career outcomes. Given that biomedical/biological engineering has more in common with biological sciences than the other engineering disciplines with which it is often included, it would be important to understand how graduate student funding relates to career outcomes for biomedical/biological engineering specifically. Similarly, postdoctoral positions in biomedical/biological engineering are also more similar in frequency and length to biological sciences, so any studies of PhD career outcomes would need to carefully consider postdoctoral employment in relation to permanent positions, and to the extent possible, career aspirations of postdoctoral researchers.

In conclusion, research at the graduate level has often considered engineering as one field of study. Our research demonstrates that when it comes to funding mechanisms for doctoral recipients, we see fairly substantial differences across engineering sub-disciplines, which has a multitude of potential influences on students’ experiences and outcomes based on prior research. Most pertinent for the readers of this journal, there are some critical differences for biomedical/biological engineering that make this important element of graduate education make the sub-discipline better resemble life sciences sub-disciplines than engineering sub-disciplines. These results further motivate the need to understand doctoral education better at the engineering sub-discipline level. Within a broader ecosystem of job markets in which biomedical engineering PhDs find employment across engineering, biotechnology, and healthcare industries, shifts in graduate student funding portfolios (e.g., fewer fellowships in favor of more research or teaching assistantships) are unlikely to impact overall employability of biomedical engineering doctorates. The problem, rather, lies in inequitable experiences and career outcomes for specific students disproportionately influenced by their funding mechanism alone. With intention, there is much advisors can do to counteract these effects and engage fellowship students in structured lab work early in their careers.