Secondary Logo

Journal Logo

An Investigation of Professional Networks and Scholarly Productivity of Early Career Physical Therapy Faculty

Becker, Betsy J. DPT, PhD; Sayles, Harlan MS; Woehler, Meredith PhD; Rost, Tony SPT; Willett, Gilbert M. PT, PhD, MS

Journal of Physical Therapy Education: June 2019 - Volume 33 - Issue 2 - p 94–102
doi: 10.1097/JTE.0000000000000094
Research Report
Free
SDC

Introduction and Review of the Literature. Physical therapy (PT) faculty must retain a scholarly agenda. Active engagement and dissemination are challenging, especially for new faculty. Evidence suggests that faculty professional networks can improve performance and innovation. The aim of this study was to determine an effective network structure and composition for scholarly activity of early career PT faculty.

Subjects. Early career faculty (less than 5 years of experience) with primary teaching and service/administrative duties who worked in accredited entry-level PT programs from institutions of varying Carnegie Classification levels. Data from 50 faculty were analyzed.

Methods. Subject questionnaires gathered data for social network analysis (visualization and calculation of network structure and composition). Participants' scholarly activity was determined by curriculum vitae analyses. Multivariable ordinary least squares regression models were developed to determine associations between networks and scholarly productivity.

Results. The results show evidence that a more open, less interconnected (ie, low density) network is associated with higher scholarly activity when controlling for the duration as a faculty member and whether the individual has an academic doctoral degree.

Discussion and Conclusion. Key implications from this study include 1) faculty can be productive in their first 5 years regardless of their institution's Carnegie Classification, days on the job, and achievement of an academic doctoral degree; 2) an effective network for scholarly productivity is one that is open and less densely interconnected; and 3) there are practical strategies faculty and their mentors can take to make networks more effective.

Betsy J. Becker is an associate professor and Program Director of the Division of Physical Therapy Education in the College of Allied Health Professions at the University of Nebraska Medical Center, 984420 Nebraska Medical Center, Omaha, NE 68198-4420 (betsyj.becker@unmc.edu). Please address all correspondence to Betsy J. Becker

Harlan Sayles is a Statistician III in the Department of Biostatistics in the College of Public Health at the University of Nebraska Medical Center.

Meredith Woehler is an assistant professor of Management in the School of Business at Portland State University.

Tony Rost is a third year Doctor of Physical Therapy student in the Division of Physical Therapy Education in the College of Allied Health Professions at the University of Nebraska Medical Center.

Gilbert Willett is an associate professor in the School of Dentistry at Creighton University.

The authors declare no conflicts of interest.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.aptaeducation.org).

Received September 05, 2018

Received in revised form December 04, 2018

Accepted December 05, 2018

Back to Top | Article Outline

INTRODUCTION AND REVIEW OF THE LITERATURE

Faculty success in higher education is achieved by optimal performance in teaching, service, clinical expertise, and scholarly productivity. Scholarly productivity is a high priority for faculty development and for promotion and tenure.1 Physical therapist (PT) faculty are required by the Commission on Accreditation in Physical Therapy Education (CAPTE) to have a scholarly agenda and to report their productivity and progress.2 Despite this requirement, “only 21% of core faculty describe themselves as actively engaged in scholarship,” as reported in the 2015 McMillian lecture by Dr. Synder-Mackler.3 A study of more than 2000 PT faculty from the 225 CAPTE-accredited entry-level physical therapy schools reported that scholarly productivity has not increased in the last 10 years, and, worse still, most PT programs reported having at least one faculty member with no distribution of scholarly work at all.4

Having and being actively involved in an effective professional network is a predictor of research publications, productivity, retention and advancement, and career satisfaction.5 There is also evidence that awareness about the potential of one's network leads to higher performance, increased innovation, and varied collaborations.5–7 In this context, the three constructs of social capital theory can explain the value of networks. The first construct is social capital that can be acquired by direct and indirect access to resources. Second, social cohesion may increase through connections with colleagues creating trust and support. Finally, social capital can improve through creating bridges across gaps in a network to facilitate the flow of information. The framework and methodology of social network analysis account for social capital through network connections, especially those with highly valued expertise.8 Faculty who are knowledgeable about their network contacts and social capital have a greater likelihood of achievement in scholarly productivity and success when in partnership with an interprofessional group.9,10

Given the relevance of scholarly activity requirements for all PT faculty by the accrediting body,2 the high numbers of new faculty joining academia,11 and the value of collaboration in scholarly work, it is important to consider network relationships that can aid in early career faculty success. However, studying how relationships contribute to an effective network for scholarship has not been applied to the field of PT, where there is a vital need to promote scholarly activity.

The aim of this study was to determine an effective network structure and composition for scholarly activity of early career PT faculty. The research question guiding this aim was, “Does the professional network of an early career PT faculty member at baseline (time 1) predict scholarly productivity 1 year later?” The 2 hypotheses were a) early career PT faculty with a professional network structure that is large with low interconnectedness (density) will have higher scholarly activity and b) early career PT faculty with a professional network composed of individuals with a variety of expertise will have higher scholarly activity.

Back to Top | Article Outline

SUBJECTS AND METHODS

Inclusion Criteria

Candidates for inclusion in the study were faculty working full-time (as defined by their institution) in an entry-level physical therapy education program (either CAPTE accredited or candidates for accreditation), in their first 5 years, with a workload that included primarily teaching and service responsibilities (40%–50% full-time equivalent [FTE]), including Directors of Clinical Education. Faculty were excluded if they had research appointments (>50% FTE of dedicated time to scholarly activity), changed institutions during the study, or had more than 1 year of full-time teaching experience in a different PT school before their current faculty appointment. This definition of early career is used by CAPTE2 and the Faculty Development committee of the Academy of Physical Therapy Education of the American Physical Therapy Association (APTA).12

Back to Top | Article Outline

Power Analysis

A power analysis determined that, with a sample of 30 participants, there would be 80% power to detect an increase in R2 of 0.15 in an ordinary least squares regression of the Scholar Score, the measure of productivity. This analysis included an adjustment for the case where a multiple regression of the Scholar Score on only a set of three control variables in the regression model would yield an R2 value of 0.3. Based on an estimated attrition rate of 40% over the yearlong study, an enrollment goal of a minimum of 42 study participants was established.

Back to Top | Article Outline

Study Participant Recruitment

The primary area of recruitment was the Academy of Physical Therapy Education of the APTA Faculty Development Workshop (FDW). Additional recruitment efforts included email announcements to the FDW attendees from the year before and postings in the electronic newsletters for the APTA Academy of Physical Therapy Education and Academy of Neurologic Physical Therapy. These sections were selected based on convenience because of professional connections established at the FDW. To reach faculty who were not members of the APTA, all PT program directors were emailed with information to forward on to eligible study participants.

Potential participants were provided a flier outlining the project, including their rights as a research participant, and were directed to the project website with a link to the survey. Participants received a $20 Amazon (Seattle, WA) gift card for each survey completed (baseline and 1-year follow-up). The participant activity through the study is shown in Figure 1.

Figure 1

Figure 1

Back to Top | Article Outline

Data Collection

Data about study participants' network structure and composition, scholarly activity, and sociodemographic information were collected through an online questionnaire, by reviewing participant curriculum vitaes (CVs), and though internet searches, as shown in Table 1.

Table 1

Table 1

Back to Top | Article Outline

Questionnaire

A 5-part online questionnaire was administered using Qualtrics software (Version 2017, Provo, UT), with an estimated completion time of 60 minutes. Part 1 included demographic items. Part 2 of the questionnaire employed an ego-network design, which asked participants (ego) for information about contacts and how they are acquainted with each other.13,14 In part 3, the network name generator section, participants listed contacts whom they considered important sources of work-related information such as teaching, scholarly activity, and service and administration. Recalling names is a known limitation10,13 and to assist in recalling network contacts, five category prompts were used, including those who work 1) in PT at the same institution as the participant; 2) in PT but at a different institution; 3) outside of PT but at the same institution; 4) outside of PT and at a different institution; and 5) primarily as a PT in clinical care. Each name entered appeared in later items, allowing for personalized questions about network contacts (eg, age, academic rank, and expertise). The fourth part was the last group of questions and included name interrelator items, which asked respondents to indicate which of their contacts know each other and could share information or ask a question. The fifth and final part was an end-of-survey message directing participants to submit their CV (Appendix A, Supplemental Digital Content 1, http://links.lww.com/JOPTE/A49).

The questionnaire was designed based on other studies of faculty networks15–17 and in consultation with experts in this field of research.18 The questionnaire was pilot tested by seven faculty with similar characteristics to the study population, with embedded open-ended questions and text boxes for comments. Modifications to the final survey included 1) minor wording changes for clarification; 2) adding the choice of Lecturer to academic rank; 3) adding Medical Doctor and Doctor of Osteopathic Medicine to categories for highest degree achieved for network contacts; 4) and embedding weblinks to scholarly activity informational videos from the project webpage as a refresher.

Back to Top | Article Outline

Calculating Network Structure and Composition Measures

Network structure and composition measures were calculated using the network analysis software program UCINET (Borgatti, S.P., Everett, M.G. and Freeman, L.C. 2002. UCINET for Windows: Software for Social Network Analysis. Harvard, MA: Analytic Technologies). First, network visualizations were generated to show the relationships between network contacts, where each dot represented an individual and the lines between them represented who was known to the other. The closer an individual appeared to another network contact on the visualization, the more relationships they share. These maps were used to visualize the general shape of a participant's network. Second, network structure measures of size and density were calculated. These included the number of persons in a network and the proportion of connections between network members (eg, how many people are connected to each other). Third, network composition measures of homophily (the similarity between the study participant and their network contacts) and heterogeneity (the diversity of network contact characteristics) were calculated. Diversity groups were also created from the single heterogeneity measures to account for the variety of characteristics and experiences a network contact brings. Detailed information about these network measures is provided in Supplemental Digital Content 2 (Appendix B, http://links.lww.com/JOPTE/A50).

Back to Top | Article Outline

Calculating Scholarly Activity

The CVs were used to gather information about scholarly activity and sociodemographic information, such as start date and specialty certifications. Successful demonstration of data collection of scholarly activity by CVs has been shown to be a noninvasive method whereby study participants are not required to re-report productivity, a potentially time-consuming activity.19 As with Halvorson et al,20 this study used both CVs and follow-up conversations with participants to clarify items.

Each CV was systematically analyzed to account for all scholarly activities using operational definitions for each category of scholarly activity. Clarifying questions were emailed to study participants when additional information was needed about a reported activity. Two evaluators, B.J.B. and G.W., discussed scholarly activities if a category was unclear and jointly finished coding CVs. Additional validation strategies included a search of library databases to confirm accuracy of the reported publications, and a review of 10% of the CVs was conducted by a third reviewer (T.R.) to compare results.

Scholarly activity counts were entered into a database and exported to statistical software (IBM SPSS Statistics for Windows, Version 24.0; IBM Corp, Armonk, NY) for calculation of a Scholar Score. The Scholar Score was a self-developed, weighted formula that accounted for the quality and quantity of presentations, publications, and grants. Program directors from the study participants' institutions (n = 39) and members of promotion and tenure committees familiar with the project (n = 6) were recruited to serve on an expert panel. Program directors have experience guiding the early career faculty including for promotion and tenure because this is a CAPTE requirement.2 Responses were received from 22 of 39 program directors and all six other committee members. Point values were determined by this panel using a reference-based scale where the reference, a peer-reviewed publication, was worth 10 points and all other activities where assigned point values accordingly. Additionally, quality was measured using a bonus for authorship order and peer-reviewed publications, grant funding amounts and external awards, presentation audience, and whether it was an invited or peer-reviewed selection process. All scholarly types described by Boyer21 were considered. The grant award amounts were selected because they are realistic for what an early career faculty may achieve. Table 2 shows the scores assigned to the study participants' reported scholarly activity items.

Table 2

Table 2

Back to Top | Article Outline

Ethics Approval

The University of Nebraska Medical Center Institutional Review Board approved this study, and all participants provided informed consent. Pseudonyms were used to protect the identity of the participants and network contacts named.

Back to Top | Article Outline

Statistical Analysis

The dependent variable accounting for scholarly activity was the Scholar Score. The independent variables were a) network structure measurements of size and density and b) network composition measures of homophily and heterogeneity. Control variables were gender, age, duration as a faculty member, academic doctorate, if the participant was a Director of Clinical Education or not, and the basic Carnegie Classification22 and public or private funding of the institution where the participant worked.

Descriptive analyses were completed to summarize the study sample. Exploratory scatter plots were generated for continuous variables, and single independent variable ordinary least squares regression models were used to determine the existence of meaningful associations between control variables, independent variables, and the primary outcome of interest, the Scholar Score. Multivariable ordinary least squares regression models were used to determine which of the independent variables were most predictive of the Scholar Score. Inclusion criteria for these models were a single predictor model P value of <.15; variables were removed from the multivariable model until all remaining P values were <.05.

Back to Top | Article Outline

RESULTS

Response Rate

At time 1 (T1), 97 responses had been received but 31 were excluded, one for not meeting the inclusion criteria and 30 for submitting an incomplete survey, leaving a total of 66 study participants. One year later at time 2 (T2), 51 of 66 possible responses (77.4%) were received. One was excluded because of a job change. Data from the remaining 50 study participants were analyzed which exceeded the minimum of 42 estimated to be needed for sufficient power to detect rather large effect sizes.

Back to Top | Article Outline

Study Participant Characteristics

Participants were 80% (n = 40) women, 96% (n = 48) White Caucasian, and most reported themselves in the age range of 35–44 years (40%, n = 20) or 45–54 years (28%, n = 14). On average, duration as a faculty member was 1.6 years ± 1.09 (range = 0.06–3.94) (Table 3). About half were on a tenure track (48%, n = 24). Most had a clinical specialty (78%, n = 39) and/or were assistant professors (90%, n = 45), but few had earned an academic doctoral degree (16%, n = 8). This sample represented 39 different institutions in 24 states. Of these institutions, about half were publicly funded (46%, n = 18). The Carnegie Classifications were doctoral (33%, n = 13), special focus (n = 6, 15%), master's (44%, n = 17), and baccalaureate (8%, n = 3). Although the response rate was high, nonresponse bias was considered. There were no statistically significant differences between respondents and nonrespondents at T2 based on gender, age, or duration as a faculty member.

Table 3

Table 3

Back to Top | Article Outline

Descriptive Statistics

Scholar Score

The mean Scholar Score at T1 was 32.7 (SD = 46.5, range 0–220) and doubled 1 year later to 66.6 (SD = 77.6, range 0–371). Although the Scholar Score doubled, this does not necessarily mean that two times more scholarly items were disseminated because the Scholar Score is weighted for quality and not just a count. Figure 2 provides an illustration of the composition of weighted items that contributed to the Scholar Score and changes that occurred over the course of the 1-year study. A statistical analysis comparing between baseline and 1 year later was not completed because scholarly productivity could only increase by virtue of the duration of time as a faculty member. Therefore, the activity at T2 has to be greater than the activity at T1; although it is not guaranteed that over a 1-year period a faculty member would disseminate a scholarly product, it is extremely unlikely that a group of 50 would not produce something.

Figure 2

Figure 2

Back to Top | Article Outline

Network Structure and Composition Measures

At T1, the mean size of the networks was 25.4 contacts (SD = 13.4, range 4–62) and mean density was 40.2% (SD = 16.6, range 18.6–100). Network shapes varied, and examples are shown in Figure 3. There were 6 measures of homophily and 18 measures of heterogeneity analyzed in this study. The analysis of diversity included a review of single measures and the aggregation of respondents into diversity groups which attempted to account for the variety of experiences and demographics of the network contacts.

Figure 3

Figure 3

Back to Top | Article Outline

Regression Analyses

Univariate Analyses

Single predictor regressions of change in Scholar Score from baseline to 1 year later are presented in Table 4. The potential control variables of race or ethnicity and academic rank were not included as there was not enough variation in the study sample. The network structure measure of density was selected for inclusion in the multivariable model (P = .067), as was the network composition measure of gender homophily (same gender as the study participant) (P = .099). No other network structure or composition measures met the P < .15 criteria. The 2 demographic measures that did meet the criteria were duration as a faculty member (P = .038) and attainment of an academic doctoral degree (P = .006).

Table 4-a

Table 4-a

Table 4-b

Table 4-b

Back to Top | Article Outline

Multivariate Analysis

A multivariable regression model was developed to determine which covariates were most predictive of the Scholar Score. Assumptions of non-multicollinearity, independence of residuals, and homoscedasticity were evaluated and no violations were found.23 The initial multivariable model with gender homophily, network density, duration as a faculty member, and attainment of academic doctorate showed that all four measures were significant predictors of Scholar Score. However, when three outlier observations were removed, gender homophily was no longer a significant predictor. Thus a decision was made to exclude it from the final model shown in Table 4. The final model showed that the network structure measure of density was a significant predictor of the Scholar Score when controlling for the duration as a faculty member and academic doctoral degree attainment (B = −1.099, P = .048).

Back to Top | Article Outline

DISCUSSION AND CONCLUSION

The study emerged from concerns related to the high number of early career faculty entering academia and the importance of supporting the evidence-based practice from which PT is based. This study used a survey to gather data to construct networks. This study was the first to use social network analysis to study network structure and the network composition of early career PT faculty. The results present new information to guide early career faculty and mentors who monitor their career advancement.

The research question guiding the study was, “Does the professional network of an early career PT faculty member at baseline (time 1) predict scholarly productivity 1 year later?” To answer this question, a longitudinal cross-sectional survey study was developed and implemented. Social network analysis was used to calculate network structure (size and density) and composition measures (homophily and heterogeneity). The results partially supported the first hypothesis where networks of early career PT faculty that are less interconnected with low density were associated with higher scholarly activity, but the size of the network was not. The second hypothesis was not supported because the results did not show association between the diversity of the network contacts' characteristics (eg, gender and expertise) and scholarly productivity. Four major points can be drawn from these findings.

Back to Top | Article Outline

Scholarly Activity

First, the Scholar Scores nearly doubled over the course of the study, indicating that 1 year was sufficient time for the study, demonstrating the correlated factor that early career faculty can be productive. Grants were the lowest category of scholarly activity. This finding is not surprising, given that it is difficult to apply for and receive a grant in the short time the faculty in this study had been employed (mean 1.6 years). Presentations were the highest category, likely because of the expedited turnaround time for acceptance compared to manuscript submissions or grant applications. The Scholar Score appeared sensitive enough to account for scholarly activity of early career PT faculty. This finding could be attributed to the inclusion of scholarship of discovery, application, teaching, and integration,21 as well as abstracts and platform presentations, where early career PT faculty are more likely to have initial success. One measure to account for all scholarly activity in the analysis versus separate counts for grants, publications, and presentations was beneficial because it is a combination of these items that contribute to career advancement.

Second, these results demonstrated that scholarly achievement is far more than just “days on the job.” Early career faculty success could be attained through intentional action toward securing an appropriate network of collaborators. It also stands to reason that if an effective network is established early in the career, higher scholarly productivity would be seen in the long term. This is an area where further investigation is warranted.

Third, CAPTE (Standard 4K) requires that 50% of the entry-level program faculty have an academic doctoral degree,24 thus, this pursuit should be encouraged for faculty to advance the PT profession. It has also been suggested that having an academic doctoral degree will lead to increased scholarly productivity in the field.3,4,25 The results of this study demonstrate that an academic doctoral degree does not guarantee scholarly productivity nor is it the only way to achieve it. Findings show new faculty who do not have an academic doctoral degree, but do have an effective network, can achieve early success even if only a small portion of their FTE is devoted to scholarly activity. Terosky and O'Meara26 reported that acting intentionally for career success with the development of a supportive network appears to serve as a mechanism for success, especially given the demanding workload of faculty members.

Back to Top | Article Outline

Professional Network

Finally, through the innovative use of social network analysis, this study found that an effective network structure for early career PT faculty is one that is less interconnected and more open. There is evidence that people who actively learn about their network can modify relationships over time versus those that are not self-aware of their actual and potential collaborators.27 Change in a network is more than increasing its size, although it is related. Early career faculty should take stock of their collaborations and the social capital gained from their network. Rodan et al6 showed profound results with business managers who modified their network from 75% of members who knew each other (highly densely interconnected) to 25% who knew each other (lower density). Another study showed that access to resources was improved when network contacts were less densely interconnected, even when these contacts were not close confidants.6 A PT faculty member could implement several practical strategies for making a network less densely interconnected.28,29

One strategy would be for the faculty member to ask mentors for introductions to others not already highly connected in their network. In this study, an average of 43% (SD 0.21) of the network contacts were identified as mentors. This indicates that early career PT faculty members have several people to reach out to for these introductions. Mentors may also initiate new introductions because this is a common activity some mentors incorporate already. New introductions increase the network size and, because these persons are not already highly interconnected, the network becomes more open. Conversely, if new members know many others within the existing network, then size increases but the network becomes more interconnected and closed, the less desirable consequence.

Attending conferences and profession-related meetings is an effective method to make new connections, especially when exploring new projects where information or skills are needed. However, it is worthwhile to note that when the event only serves to facilitate meeting new people, results are less positive. In one study, persons interested in meeting new connections attended a networking event to facilitate introductions, but with no incentive such as an upcoming project in which to express interest or converse, participants gravitated to known friends and network interconnectedness changed very little.30

Preventing a network from becoming too densely interconnected includes a combination of networking through both formal groups (eg, professional organizations and journal clubs) and less formal interactions, such as talking to an unknown person sitting beside one on an airplane.31 This chance meeting can develop into relationships of value given the right situation,32 and therefore an effective third strategy for early career PT faculty is to get involved in different activities to make new contacts especially with people unlike themselves.

Using one or more of these strategies could allow early career faculty members to develop less densely interconnected networks and gain the social capital of novel information, opportunities, and resources to create an effective network for scholarly productivity. The innovation process for creating scholarly activity within an organization is influenced by the relationships among network contacts. Interacting with these persons and learning from them is a critical component of this process.33 Network analysis was a successful methodology for analyzing the support system and collection of potential and actual collaborators for early career faculty. The analysis made these connections visible for the early career faculty who reside at the lower end of the academic hierarchy in terms of tenure, academic rank, and scholarly productivity. Their social capital could be very high if the network member knowledge and expertise are used to build scholarly agendas.17,34

Back to Top | Article Outline

Limitations

This study is not without limitations. The design included an in-depth review of the individual professional network; however, there are many other factors that may have an impact upon productive scholarly activity from an organizational level, including the type of institution, organizational structure, and productivity of other faculty members. Individual factors must also be considered, including personal and family obligations outside of work. Analysis of the networks of individual respondents, using an ego-network design, can also only account for relationships formed without consideration of the pool from which members have been selected. Therefore, it is impossible to account for errors of omission when a network member is not listed, or errors of commission when a network member is included but that relationship nevertheless does not exist. Because the ego-network design includes perceptions of network connections among contacts, it is impossible to determine the existence of connections reported. Thus, it is possible that a study participant was unaware of a potential connection.

It should also be noted that scholarly activity is a long process and the length of this study, 1 year, is a short time to observe or account for it. Attempts to mitigate this limitation included accounting for the more preliminary dissemination of scholarly work, such as abstracts and platform presentations, and using the Scholar Score.

The results did not show significant findings related to the network composition measure of diversity of members' characteristics, heterogeneity (eg, gender, age, academic rank, and expertise). It is possible that the sample size was not large enough nor did it have sufficient variability in network composition to detect significant effects. In this study, respondents were to identify network members who could help with scholarly activity. Therefore, it is likely some identified will not be actual collaborators but may be valuable for other aspects of work not studied here.

Back to Top | Article Outline

Future Study Plans

Plans to continue exploring the relationships between the professional networks and scholarly activity of PT faculty include analyzing subsets of the networks studied here, such as mentor-only networks or networks of women faculty. It would also be interesting to explore the process of how the network is developed and what constitutes a valuable network member through a qualitative grounded theory study.

Back to Top | Article Outline

CONCLUSION

The study was grounded in the principle that social capital can be gained from an effective professional network to aid in faculty success. The starting point of this project identified a knowledge gap in the field of early career PT faculty who are developing a scholarly agenda with the engagement of a professional network. High numbers of new faculty are projected to enter the academy, many without in-depth training in traditional research through the requirements of obtaining an academic doctoral degree and many experiencing for the first time the requirements of and finite time to fulfill the academic obligations of teaching, service, clinical work, and scholarly activity. The results of this study show that the most effective network to enable higher Scholar Scores is one that is less densely interconnected. This research will contribute to the literature by identifying an effective network and proposing strategies for network development. Regardless of early career PT faculty academic credentials, strategies for creating an effective network of collaborators should be implemented to aid in career advancement in the area of scholarly activity.

Back to Top | Article Outline

Acknowledgments

The authors would like to thank Susanna Von Essen, MD, MPH, of the University of Nebraska Medical Center (UNMC) for her guidance, commitment to high standards, and insight throughout the study. In addition, UNMC faculty members Laura Bilek, PT, PhD, Victoria Kennel, PhD and Christopher J. Kratochvil, MD also provided valuable feedback.

Back to Top | Article Outline

REFERENCES

1. Kaufman RR. Career factors help predict productivity in scholarship among faculty members in physical therapist education programs. Physical TherapyPhys Ther. 2009;89:204–216.
2. Commission on Accreditation for Physical Therapy Education (CAPTE) standards and required elements for accreditation of physical therapist education programs. http://www.capteonline.org/uploadedFiles/CAPTEorg/About_CAPTE/Resources/Accreditation_Handbook/CAPTE_PTStandardsEvidence.pdf. Accessed December 7, 2017.
3. Snyder-Mackler L. Mary McMillan lecture: Not eureka. Phys Ther. 2015;95:1446–1456.
4. Hinman MR, Brown T. Changing profile of the physical therapy professoriate—Are we meeting CAPTE's expectations? J Phys Ther Educ. 2017;31:95–104.
5. Warner ET, Carapinha R, Weber GM, Hill EV, Reede JY. Faculty promotion and attrition: The importance of coauthor network reach at an academic medical center. J Gen Intern Med. 2016;31:60–67.
6. Rodan S, Galunic C. More than network structure: How knowledge heterogeneity influences managerial performance and innovativeness. Strateg Manag J. 2004;25:541–562.
7. Law M, Wright S, Mylopoulos M. Exploring community faculty members' engagement in educational scholarship. Can Fam Physician. 2016;62:e524–e530.
8. Crossley N, Bellotti D, Edwards G, Everett MG, Koskinen J, Tranmer M. Social capital and small worlds: A primer. In: Social Network Analysis for Ego-Nets. Thousand Oaks, CA: Sage; 2015:23–43.
9. Ponjuan L, Conley VM, Trower C. Career stage differences in pre-tenure track faculty perceptions of professional and personal relationships with colleagues. J Higher Education. 2011;82:319–346.
10. Borgatti SP, Everett MF, Johnson JC. Analyzing Social Networks. Los Angeles, CA: Sage; 2013.
11. CAPTE aggregate program data: Physical therapist education fact sheets 2016-2017. http://www.capteonline.org/uploadedFiles/CAPTEorg/About_CAPTE/Resources/Aggregate_Program_Data/AggregateProgramData_PTPrograms.pdf. Accessed May 7, 2018.
12. American Physical Therapy Association faculty development. http://aptaeducation.org/events/faculty-development-workshop/2018/. Accessed April 25, 2018.
13. Yenigün D, Ertan G, Siciliano M. Omission and commission errors in network cognition and network estimation using ROC curve. Social Networks. 2017;50:26–34.
14. Borgatti SP, Everett MG, Johnson JC. Data collection. In: Analyzing Social Networks. Thousand Oaks, CA: Sage; 2013:44–61.
15. Anderson MH. Social networks and the cognitive motivation to realize network opportunities: A study of managers' information gathering behaviors. J Organizational Behav. 2008;29:51–78.
16. Jippes E, Achterkamp MC, Brand PLP, Kiewiet DJ, Pols J, van Engelen JML. Disseminating educational innovations in health care practice: Training versus social networks. Soc Sci Med. 2010;70:1509–1517.
17. Niehaus E, Meara K. Invisible but essential: The role of professional networks in promoting faculty agency in career advancement. Innovative Higher Educ. 2015;40:159–171.
18. Univeristy of Kentucky Gatton College of Business and Economics. Social network anlaysis. https://sites.google.com/site/linkscenterworkshop2016/. Accessed June 6, 2016.
19. Finney JW, Amundson EO, Bi X, et al. Evaluating the productivity of VA, NIH, and AHRQ health services research career development awardees. Acad Med. 2016;91:563–569.
20. Halvorson MA, Finlay AK, Cronkite RC, et al. Ten-year publication trajectories of health services research career development award recipients: Collaboration, awardee characteristics, and productivity correlates. Eval Health Professions. 2016;39:49–64.
21. Boyer EL. The scholarship of teaching from scholarship reconsidered: Priorities of the professoriate. Coll Teach. 1991;39:11.
22. University ToI. Carnegie classification of institutions. http://carnegieclassifications.iu.edu/classification_descriptions/basic.php. Accessed March 29, 2018.
23. Vittinghoff E, Glidden D, Shiboski S, McCulloch C. Regression Methods in Biostatistics: Linear, Logistic, Survival, and Repeated Measures Models. 2nd ed. New York, NY: Springer-Verlag; 2012.
24. CAPTE. Standards and required elements for accreditation of physical therapist education programs. http://www.capteonline.org/uploadedFiles/CAPTEorg/About_CAPTE/Resources/Accreditation_Handbook/CAPTE_PTStandardsEvidence.pdf. Accessed August 1, 2018.
25. Bliss R, Brueilly KE, Swiggum MS, Morris GS, Williamson EM. Importance of terminal academic degreed core faculty in physical therapist education. J Phys Ther Educ. 2018;32:123–127.
26. Terosky AL, O'Meara K. The power of strategy and networks in the professional lives of faculty. Liberal Education. 2011;97:54–59.
27. Burt RS, Ronchi D. Teaching executives to see social capital: Results from a field experiment. Soc Sci Res. 2007;36:1156–1183.
28. Burt Ronald S. Structural holes and good ideas. Am J Sociol. 2004;110:349–399.
29. Uzzi B. Keys to understanding your social capital. J Microfinance/ESR Rev. 2008;10:3.
30. Burkus D. Go ahead, skip that networking event. Harv Business Rev. 2018:3. https://hbr.org/2018/05/go-ahead-skip-that-networking-event. Accessed May 14, 2018.
31. Feld SL. The focused organization of social ties. Am J Sociol. 1981;86:1015–1035.
32. Ryan L. Looking for weak ties: Using a mixed methods approach to capture elusive connections. Sociological Rev. 2016;64:951–969.
33. Abbasi A, Altmann J. On the correlation between research performance and social network analysis measures applied to research collaboration networks. In Proceedings from the 44th Hawaii International Conference on System Sciences; January 4-7, 2011; Kauai, HI.
34. Cross R, Borgatti SP, Parker A. Making invisible work visible: Using social network analysis to support strategic collaboration. Calif Manag Rev. 2002;44:25–46.
35. Crossley N, Bellotti D, Edwards G, Everett MG, Koskinen J, Tranmer M. Analyzing ego-net data. In: Social Network Analysis for Ego-Nets. Thousand Oaks, CA: Sage; 2015:76–104.
    Keywords:

    Faculty development; Network analysis; Scholarly activity; Physical therapy

    Supplemental Digital Content

    Back to Top | Article Outline
    Copyright 2019 © Academy of Physical Theraphy Education