Eurasian Bus Rev https://doi.org/10.1007/s40821-018-0102-3 ORIGINAL PAPER

Interdisciplinarity: who reaps the benefits? Thomas Grebel1 • Uwe Cantner2,3 • Julia Schumm4

Received: 23 May 2017 / Revised: 18 December 2017 / Accepted: 22 December 2017 Ó Eurasia Business and Economics Society 2018

Abstract Interdisciplinary research has become increasingly popular in medical science. It offers a large potential for new perspectives, new inventions, and for researchers more opportunities to publish. Whether all researchers benefit equally from this opportunity when doing interdisciplinary research is unclear. Using data from a survey on German researchers in medical clinics and institutes, we investigate the determinants of research success measured by publication. We control for organizational differences and researchers’ experience level. Running negative binomial regressions, the results suggest that interdisciplinary research is beneficial for researchers on the executive level at institutes with little patient care. On lower hierarchical levels, interdisciplinary research contributes less to publication performance. Keywords Knowledge creation  Basic research  Interdisciplinarity  Research incentives  Empirical research JEL Classification I00  O31  D01

& Thomas Grebel [email protected] 1

Economics Department, TU Ilmenau, Ehrenbergstr. 29, 98684 Ilmenau, Germany

2

Economics Department, University of Jena, Carl-Zeiß-Str. 3, Jena, Germany

3

Department of Marketing and Management, University of Southern Denmark, Odense, Denmark

4

Dr. Horst Schmidt Klinik, Department of Cardiology, Wiesbaden, Germany

123

Eurasian Bus Rev

1 Introduction Interdisciplinary research is on the rise (Van Noorden 2015) and has become one of the main drivers of knowledge production (Gibbons et al. 1994). In general, it is considered to be beneficial (Carayol and Thi 2005; Nissani 1997), particularly in medical science, where multifactorial disorders such as AIDS, cancer, or cardiovascular indispositions require interdisciplinary approaches. According to the survey by Rhoten and Parker (2004), the majority of researchers (72% of respondents) across all tiers of the academic personnel consider interdisciplinary research beneficial to their career. One of the main benefits, besides having discovered something new, is that they publish and thus follow the inevitable imperative of modern sciences: publish or perish (Guraya et al. 2016). Researchers’ personal publication record is constitutive for finding jobs, receiving promotions or research funding (Fanelli 2010). The results by Rhoten and Parker (2004) are surprising, because the benefits, in terms of publication output, are not evenly distributed. Neither do all disciplines nor all researchers benefit from this trend. Clinical studies in medicine, for example, tend to be far less interdisciplinary than social studies of medicine (Van Noorden 2015). The reasons may be plentiful. It might be the speciality of a discipline (Van Noorden 2015) or inadequate funding structures that inhibit interdisciplinary research (Viseu 2015). Even when ignoring these institutional factors, the bias in publication output among researchers should still remain. Researchers work in different organizations, different hierarchies, and have different experience levels and all these circumstances influences the scope for doing research and the chance to perform interdisciplinary research. The aim of this paper is to investigate the benefits for individual researchers in terms of publication output, asking whether all researches within the field of heart medicine benefit equally. Looking at this specialized field, we investigate the incentive structure for medical researchers and search for possible factors which may influence researchers’ scope for interdisciplinary research. The data we use was collected in a survey conducted by Cantner et al. (2009) in 2008, when about 18,000 medical researchers in heart medicine in Germany were addressed via an online questionnaire to inquire about the organizational structure and their research output. Respondents were asked about their qualification profile, work load, networking activities, research autonomy, emergence of research groups and the degree of interdisciplinarity in their research. The respondents belong to different hierarchical levels (assistant, attending physician, head of department) and stem from different medical organizations (theoretical institutes, clinical institutes with or without beds). Using count-data regression we estimate the impact of interdisciplinary research on publication output on the individual level. The main results of our study show that interdisciplinary research has a positive effect on researchers’ publication output, but with a bias toward researchers on higher hierarchical levels. For researchers on lower levels, interdisciplinary research contributes less to publication success.

123

Eurasian Bus Rev

Before explicating this exercise, we motivate the need for, the incentives, and the kind of inhibitors associated with interdisciplinary research in Sect. 2. Section 3 describes the database, results are presented in Sects. 4, 5 delivers a discussion, and Sect. 6 concludes.

2 Benefits and inhibiting factors to interdisciplinary research The awareness of the need for interdisciplinary research has been steadily increasing in recent years, also in medical science (Van Noorden 2015). In spite of the obvious benefits of interdisciplinary research, there still are researchers that tend to ‘‘apply disciplined research approaches to undisciplined problems’’.1 This is due to quite many inhibiting institutional factors: medical clinics and institutes are predominantly organized in line with traditional disciplines, funding sources tend to allocate their funds according to traditional disciplines,2 and universities hesitate to educate students in interdisciplinary programs being worried about graduates’ job prospects (Bruhn 1995). The collegiate-reputation-based reward system, which is prevalent in academia (Dasgupta and David 1994), exacerbates this problem even more. The peer-review process is as much aligned to traditional disciplines as scientific organizations (Carayol and Thi 2005). Hence, also the peer-review process favors disciplinary research work. There are far less outlets for interdisciplinary research work and much less qualified researchers able to review interdisciplinary work (Choi and Pak 2007). So it comes with little surprise that researchers follow a disciplinary approach to safeguard their career. Otherwise, they may run higher risks in finding a job, because traditional academic degrees enjoy a higher reputation (Weingart and Stehr 2000). Last not least, it is difficult to master one discipline and to master more than one seems futile (Choi and Pak 2007).3 Obviously, there are pros and cons to perform interdisciplinary research. The objective here is not to investigate the validity of promoting or inhibiting factors but to look at the distribution of benefits among researchers measured by publication output. Our data contains information about German medical organizations and the researchers’ scope for (interdisciplinary) research in heart medicine. As the survey concentrates on a single, specialized field, we may assume that researchers enjoy the same context as far as the collegiate-reputation-based reward system is concerned.4 Thus we control to some extent for unobserved heterogeneity. 1

Rose (1986) as cited in Bruhn (1995, p. 332).

2

See Bruhn (1995) who describes the case of the National Institute of Health (NIH).

3

Garwin (1995) gives an insightful example. She compares researchers with musicians. A musician should concentrate on a single instrument than on many in order to excel. For both the musician and the researcher, it is rational to specialize. However, at the same time both should open up: musicians in an orchestra, because they need to be sympathetic to each other to make beautiful music, and researchers, because who else should solve pressing problems that require interdisciplinary approaches.

4

Heart medicine comprises all diseases which involve the heart and blood vessels. For a quick overview see e. g. the American Heart Association: http://www.heart.org/HEARTORG/.

123

Eurasian Bus Rev

Before looking at the role of interdisciplinary research in publication output, we have to account for organizational differences within the field of heart medicine. The three types of organizations identified below primarily differ in the work load in patient care. The scope to perform research in clinics with beds is more restricted than in theoretical clinics without beds, that is, clinics with a clear focus on research. This leads us to the first hypothesis: Hypothesis 1: Physicians’ research space and their opportunity to perform interdisciplinary research depends on the type of organization: The higher the work load of patient care, the lower physicians’ research space and scope for interdisciplinary research. Given a certain degree of research space and the opportunity to perform disciplinary research, more experienced researchers should be more effective. They should manage to publish more than younger colleagues and with respect to interdisciplinary research, they should be more capable of combining different disciplines, which should increase their publication output in the end. The respective hypothesis is the following: Hypothesis 2: Physicians’ publication output is positively related to their research space and the opportunity to perform interdisciplinary research. (a) (b)

More experienced physicians show a higher publication performance. Physicians on higher hierarchy levels are more effective in exploiting their research space and the opportunities of interdisciplinary research.

3 Data The data we use stem from a survey conducted in 2008 during a research project at the Friedrich-Schiller University in Jena. It gathers information about the organizational structure of research activities at medical clinics and institutes.5 More than 18,000 researchers in heart medicine all over Germany were addressed by an online questionnaire about personal research environments. We identified all medical researchers in cardiovascular science by their website, theoretical institues (INST1), clinical institutes without beds (INST2) and clinical institutes with beds (INST3), and addressed all researchers of these organizations via e-mail. With 1747 responses, the response rate amounts to 10%. A higher rate would certainly be preferable, but for confidentiality reasons, it was prohibited to capture the individuals identity and readdress respondents. Nevertheless, we could retrieve that the survey is representative to the extent that (1) respondents stem from all organizations doing research in heart medicine in Germany and (2) the distribution of respondents across types of institute and hierarchy level reflects the expected distribution. Hence, the survey reflects the population of cardiovascular researchers in Germany, though responses may suffer from a self-selection bias. 5

Compare Cantner et al. (2009, chapter 7).

123

Eurasian Bus Rev

Respondents were questioned about their qualification profile, work load, networking activities, research autonomy, emergence of research groups and the degree of interdisciplinarity in their research.6 All three types of organizations, theoretical institutes (INST1), clinical institutes without beds (INST2) and clinical institutes with beds (INST3), engage in research, the main difference lies in the work load in patient care. Researchers in INST3 organizations bear the highest load of patient, followed by clinics without beds, and theoretical institutes. This variable will be used to account for organizational differences. After cleaning the dataset for missing values, we count 1636 returned questionnaires, 258 stem from theoretical institutes, 288 from clinical institutes without beds and the majority of 1090 from clinical institutes with beds. This may quite well represent the organizational structure of medical research in Germany. The physicians’ position in the hierarchy may serve as a proxy for their experience level. There are four different hierarchies: research assistants at the lowest level, researchers and attending physicians at the middle level, researchers as head of department, and directors of the institute. Researchers at the director level have reached the final career stage. Attending physicians and head of departments belong to the group of researchers that explicitly pursue an academic career, whereas the majority of assistant researchers become medical practitioners not pursuing an academic career and hence finish their university degree with a doctoral degree. Therefore we build three categories of researchers: HA assistant researchers, HM mid-level faculty researchers pursuing an academic career, and the group of directors HP , who have reached full professorship. Table 1 summarizes the variables with reference to the underlying question in the questionnaire (see appendix). Question 1 and 2, concerning the type of institute and the level of hierarchy, respectively, were translated into dummy variables. Question 3 is based on a 7-Likert scale and serves as proxy for interdisciplinarity (Idis). We rescale the variable to 0 (= mainly disciplinary research), 1 (= balance between disciplinary and interdisciplinary research), and 2 (= mainly interdisciplinary research), where the Likert scale values are redistributed as in the following: {1, 2}!0, {3, 4, 5}!1, and {6, 7}!2. Question 4 and 5 are contracted to a latent variable called Rspace, which expresses a physicians freedom to do research. Question 6 gives the self-reported average number of publications per year, NumPub. The summary statistics are gathered in Table 2. The number of annual average publications NumPub is right-skewed. On average, respondents have less than four publications a year. This means that few physicians have a high (self-reported) number of publications. The variable Idis, which proxies the degree of interdisciplinary research, shows that, on average (4.6), interdisciplinary research groups seem not to be the dominant organizational form of research within a given institute. The variable Rspace indicates the physicians’ freedom to engage in research. A higher load of patient care leads to less available space to perform research. This variable is constructed via a factor analysis contracting two questions (questions 4 and 5) to one latent variable. From the mean values of the hierarchy variables (HA 6

See Cantner et al. (2009) for the questionnaire.

123

Eurasian Bus Rev Table 1 Variables/description Variable

Description

INST1

Theoretical institute

1

Dichotomous

INST2

Clinical institute without beds

1

Dichotomous

INST3

Clinical institute with beds

1

Dichotomous

HA

Researchers (assistants)

2

Dichotomous

HM

Attending physician, head of department

2

Dichotomous

HP

Full professor (director)

2

Dichotomous

Idis

Interdisciplinary teams exist

3

7-Likert

Rspace

Organizational slack for doing research

4, 5

Real (latent variable)

NumPub

Average ] of yearly publications

6

Integer

Table 2 Descriptive statistics

See description in Table 1

Question

Unit

Variable

Mean

Std. Dev.

Min.

Max.

N

NumPub

3.396

4.632

0

88

1636

Idis

4.582

1.887

1

7

1636

Rspace

0.004

0.507

- 0.536

1.057

1636

HA

0.557

0.497

0

1

1636

HM

0.388

0.487

0

1

1636

HP

0.055

0.228

0

1

1636

INST1

0.164

0.371

0

1

1636

INST2

0.176

0.381

0

1

1636

INST3

0.659

0.474

0

1

1636

HP ) we can infer that most researchers in the sample belong to lower hierarchies. Similarly, the mean values of the organization-type (INST1-INST3) show that the majority of physicians work at clinics with beds (INST3). In other words, most physicians belong to the group of young physicians (HA ) at clinics with beds (INST3) which mirrors the expected frequency distribution with the majority of researchers being trained as medical practitioner. Hence, for most respondents patient care plays a dominant role in their daily business. Table 3 provides pairwise correlations.

4 Results In the following, we test hypotheses 1 and 2. The results in Subsect. 4.1 corroborate the hypothesis that the type of organization physicians work in has a significant impact on the scope of research and the degree of interdisciplinarity. The results in Subsect. 4.2 suggest that it is senior researchers that benefit most from interdisciplinary research in terms of publication output.

123

Eurasian Bus Rev Table 3 Correlation table Variables

(1)

(2)

(3)

(1)

NumPub

1.00

(2)

Rspace

0.06*

1.00

(3)

Idis

0.19*

0.12*

1.00

(4)

(5)

(6)

(7)

(8)

(4)

HA

- 0.35*

0.14*

-0.11*

1.00

(5)

HM

0.24*

-0.14*

0.07

-0.90*

1.00

(6)

HP

0.26*

-0.01

0.08*

-0.27*

-0.19*

(7)

INST1

0.02

0.42*

0.05*

0.05

-0.05*

0.00

1.00

(8)

INST2

-0.03

0.15*

0.00

0.03

-0.05*

0.04

-0.20*

1.00

(9)

INST3

0.01

-0.45*

-0.04*

-0.07*

0.08*

-0.03

-0.61*

-0.65*

(9)

1.00

1.00

5% significance level

4.1 Organizational effects To test hypothesis 1 we regress Idis and Rspace on variables INST1-INST3, respectively. Table 4 shows the results. Model (1) is a standard OLS regression. The base category of the dummy variables is INST1. INST2 is negative and significant at the 10% level. This means that research groups in clinics without beds tend to be less interdisciplinary than those at theoretical institutes. Looking at the coefficient of INST3, we observe that the physicians indicate an even lower degree of interdisciplinary research groups in their organizations. As Idis is measured on a Likert scale, the outcome values are no continuous numbers, OLS regression can only serve as a first guess. Using the correct model, i.e. an ordered probit model, this is taken into account. Model (2) presents the corresponding results and supports the estimate of model (1): the more patient care, the less interdisciplinary research groups. The majority of respondents indicate that they perform their research autonomously, independent from others. The available space to do research, as model (3) shows, declines from INST1 to INST3. Physicians at theoretical institutes Table 4 Regressions of Idis and Rspace

Variables

INST2 INST3 Constant

(1) OLS Idis

(2) Ord. prob Idis

(3) OLS Rspace - 0.644***

- 0.231*

- 0.127*

(0.161)

(0.089)

(0.075)

- 0.519***

- 0.302***

- 1.301***

(0.130)

(0.072)

4.953*** (0.117)

Base category: INST1

Observations

1636

Standard errors in parentheses

R-squared

0.011

***p \ 0.01, **p \ 0.05, *p \ 0.1

LL

- 3350

(0.061) 0.980*** (0.055)

1636

1636 0.239

- 3055

- 2102

123

Eurasian Bus Rev

(INST1) indicate to have more space for research than physicians at clinics without beds (INST2), followed by clinics with beds (INST3). Hence, the results corroborate hypothesis 1. Physicians’ research space and their opportunity to perform interdisciplinary research depend on the type of organization: The higher the load of patient care (going from INST1 lowest to INST3 highest), the lower physicians’ research space and scope for interdisciplinary research. 4.2 Individual effects Concerning the question who reaps most of the benefits from interdisciplinary research in terms of publication, we take a look at the individual factors that may influence the probability of researchers to publish. As dependent variable, we use the number of publications (NumPub).7 As this variable is count data and overdispersed, we use negative binomial regression. Table 5 presents five models, by which we sequentially introduce the independent variables in order to show the sensitivity of the coefficients with respect to changes in coefficients. Model (4) includes Rspace and Idis as explanatory variable. Both variables have a positive impact on publication success. This holds also for models (5), (6), (7), and (8). The more research space researchers have, the higher the average self-reported annual number of publications. Also interdisciplinary research contributes to publication success in all four models. As we do not have more information about individual factors, the hierarchical position of a researcher serves as a proxy for a researcher’s experience.8 At each hierarchy level, the qualification requirements change. Consequently, researchers’ educational profile, their work and research experience increase with hierarchy. Model (5) indicates a positive relationship between hierarchy and publication success. Compared to young, less experienced researchers (HA ), researchers at hierarchy level HM publish 0.88 and at HP 1.38 papers more. Hence, being full professor, i.e. director of a clinic, interdisciplinary research appears to offer better opportunities to publish. To test hypothesis 2, we need to disentangle the individual effects of researchers’ experience in combination with Rspace and Idis. Firstly, we interact the variable Idis with hierarchy. Model (6) lays out the respective results. Again, the coefficients of Rspace and Idis remain positive and significant. Combined with interdisciplinary research, as the coefficients of HM  Idis and HP  Idis suggest, higher tiers manage to publish disproportionately more than disciplinary researchers at the same hierarchy level. Secondly, interacting hierarchy with Rspace a similar picture emerges (model 7). The coefficients of the interaction terms of Rspace with hierarchy take a positive sign. The indirect effect of Rspace adds about 0.09 publications at level HM and 0.27 at level HP . Hence, researchers on higher hierarchy levels exploit available 7

Endogeneity of self-reported data will be discussed in Section (V).

8

Unfortunately, we could not retrieve further information on the individual level due to confidentiality restrictions.

123

Eurasian Bus Rev Table 5 Negative binomial regression: dependent variable NumPub Model Idis Rspace

(4)

(5)

(6)

(7)

0.311***

0.206***

0.106**

0.204***

0.115***

(0.032)

(0.029)

(0.042)

(0.029)

(0.042)

0.047*

0.109***

0.110***

0.062**

0.072**

(0.024) HM HP

(0.022)

(0.022)

(0.029)

(0.029)

0.881***

0.664***

0.877***

0.679*** (0.086)

(0.044)

(0.085)

(0.044)

1.382***

1.074***

1.360***

1.105***

(0.081)

(0.180)

(0.081)

(0.181)

HM  Idis

0.176***

HP  Idis

0.161***

(0.059)

(0.060)

0.229**

0.193*

(0.113) HM  Rspace HP  Rspace Constant

(8)

(0.114) 0.085*

0.063

(0.044)

(0.045)

0.272***

0.251**

(0.103)

(0.104)

0.818***

0.411***

0.528***

0.421***

0.524***

(0.046)

(0.046)

(0.058)

(0.046)

(0.058)

Observations

1636

1636

1636

1636

1636

LL

- 3745

- 3434

- 3428

- 3424

- 3420

Standard errors in parentheses ***p \ 0.01, **p \ 0.05, *p \ 0.1

research space more effectively. When adding a full set of interaction terms, these results remain robust. Model (8) illustrates that the direct effects of Idis and Rspace remain positive as well as their indirect effects, although they lose in significance. The coefficient of HM  Rspace is no longer significant.

5 Discussion This empirical exercise shows that interdisciplinarity in heart medicine is conducive to researchers’ publication output. Researchers at theoretical institutes (INST1) report more interdisciplinary research than clinics without beds (INST2), followed by clinics with beds (INST3). This also reflects the order of work load in patient care, since researchers in INST3 indicate the highest work load among all types of organizations. More experienced researchers, i.e. researchers on higher hierarchy levels, are more effective in exploiting the opportunities of interdisciplinary research in terms of publication output. They also appear to be more effective in exploiting their available research space in comparison to less experienced researchers. This may be attributed to a higher experience level but could also be inherent to the system of

123

Eurasian Bus Rev

research in medical science, as senior researchers tentatively have the power to raise a claim on researchers’ publication output, as lower-level researchers are subject to directives. A caveat, we could not control for, is the possibility that researchers’ responses might be biased. Self-reported data, as Chirikos and Nestel (1984); Anderson and Burkhauser (1985) or Bazzoli (1985) argue, is subject to systematic errors. They compare self-reported health measures with other more objective health measures. The upward bias they find, however, is quite dispersed. Such a bias may also be inherent to self-reported publication records, whereby it is unclear, which direction this bias takes. For identification, we would need to compare self-reported publication output data with actual publication records—a task, we could not perform due to privacy restrictions. Conversely, the counter argument is also maintained in literature. Haveman et al. (1989); Stern (1989) or Mossey and Shapiro (1982) put forward that self-reported health measures are reliable. As a matter of fact, more recent studies, such as in Dwyer and Mitchell (1999), point out that despite the limitations, self-reported data are mostly treated as an objective proxy for health. Whether an analogy to selfreported publication records can be drawn, we could not verify. The fact that the average number of publications is no subjective sentiment but a verifiable measure in contrast to self-reported health measures, makes us believe that self-reported publication data should be a good proxy for actual publication output and that the propositions of this study are robust.

6 Conclusion There are some insights, we may gain from this exercise. The available space to do research seems not to be the sole driver of publication success as far as it concerns the discipline of heart medicine. Also interdisciplinary research appears to have a significantly positive effect on publication output. This effect, however, is moderated by institutional and individual factors, which is subsumed under the type of organization and the level of hierarchy in our paper. Researchers on higher hierarchical levels from institutes with a lower load of patient care benefit most from interdisciplinary research. For researchers at lower hierarchy levels, i.e. younger researchers from clinics with a high load of patient care, this effect is the lowest. The shortcomings in the study are primarily due to the underlying privacy restrictions. There are a lot more conceivable variables which influence researchers’ publication success. The survey data, however, does not cover further information on individual characteristics, which could deliver additional information on the drivers of publication success. Available information on formal educational levels turned out to be useless. Educational levels correlate significantly with hierarchical levels, which causes multicollinearity problems. Information on researchers location and the reputation of their research institute are not available, either. Privacy restrictions also prevented tackling the possible bias in self-reported data, which could be easily corrected. Comparing self-reported publication records with

123

Eurasian Bus Rev

actual records would have revealed any bias. On the other hand, relaxing the privacy conditions would presumably reduce the willingness of researchers to participate in such a survey. An update of the survey would also be of interest. Some institutional aspects have possibly changed over time, whereas the fundament of the collegiate-reputationbased reward system has not. The peer-review process still decides about publication success. It takes a significant influence on researchers’ academic career, as any opportunity in research, which is conducive to publication output, sets an incentive for researchers to take the chance, whereas any action reducing the propensity to publish will simultaneously reduce the incentive. Conclusively, we need to monitor and question the incentive system in academia from time to time and ask whether it still is in line with society’s needs.

Appendix See Table 6. Table 6 Questionnaire No.

Question

Measure

1.

At what type of institute do you work?

INST type 1, 2, or 3

2.

What is your hierarchy level?

Hierarchy 1, 2, 3

3.

Research groups are formed among different disciplines

7-item Likert scale

4.

Patient care is dominant in my work, which has a significantly negative impact on my research work

7-item Likert scale

5.

To reflect on research questions does not fit into my work schedule

7-item Likert scale

6.

On these grounds, I manage to publish ____ publications per year

number of publications

References Anderson, K. H., & Burkhauser, R. V. (1985). The retirement-health nexus: A new measure of an old puzzle. Journal of Human Resources, 20(3), 315–330. Bazzoli, G. J. (1985). The early retirement decision: New empirical evidence on the influence of health. Journal of Human Resources, 20(2), 214–234. Bruhn, J. (1995). Beyond discipline: Creating a culture for interdisciplinary research. Integrative Physiological and Behavioral Science, 30(4), 331–341. Cantner, U., Ferrari, M., Figulla, H.-R., Grebel, T., Hartz, S., Reich, R., Schlosser, M., Schumm, J., & Wilfer, T. (2009). Chancen und Risiken minimalinvasiver, kardiochirurgischer Herzmedizin. Bundesministerium fu¨r Bildung und Forschung.

123

Eurasian Bus Rev Carayol, N., & Thi, T. U. N. (2005). Why do academic scientists engage in interdisciplinary research? Research Evaluation, 14(1), 70–79. Chirikos, T. N., & Nestel, G. (1984). Economic determinants and consequences of self-reported work disability. Journal of Health Economics, 3(2), 117–136. Choi, B. C. K., & Pak, A. W. P. (2007). Multidisciplinarity, interdisciplinarity, and transdisciplinarity in health research, services, education and policy: 2. Promotors, barriers, and strategies of enhancement. Clinical and Investigative Medicine, 30(6), E224–E232. Dasgupta, P., & David, P. (1994). Toward a new economics of science. Research Policy, 23(5), 487–521. Dwyer, D. S., & Mitchell, O. S. (1999). Health problems as determinants of retirement: Are self-rated measures endogenous? Journal of Health Economics, 18(2), 173–193. Fanelli, D. (2010). Do pressures to publish increase scientists’ bias? An empirical support from US States data. PloS One, 5(4), e10271. Garwin, L. (1995). In praise of interdisciplinarity. Nature, 376(6541), 547–547. Gibbons, M., Limoges, C., Nowotny, H., Schwartzman, S., Scott, P., & Trow, M. (1994). The new production of knowledge: The dynamics of science and research in contemporary societies. London: Sage. Guraya, S. Y., Norman, R. I., Khoshhal, K. I., Guraya, S. S., & Forgione, A. (2016). Publish or perish mantra in the medical field: A systematic review of the reasons, consequences and remedies. Pakistan Journal of Medical Sciences, 32(6), 1562. Haveman, R., Wolfe, B.L., & Huang, F.M. (1989). Disability status as an unobservable: estimates from a structural model. National Bureau of Economic Research, Working Paper No. 2831. Mossey, J. M., & Shapiro, E. (1982). Self-rated health: A predictor of mortality among the elderly. American Journal of Public Health, 72(8), 800–808. Nissani, M. (1997). Ten cheers for interdisciplinarity: The case for interdisciplinary knowledge and research. The Social Science Journal, 34(2), 201–216. Rhoten, D., & Parker, A. (2004). Risks and rewards of an interdisciplinary research path. Science, 306(5704), 2046. Rose, R. (1986). Disciplined research and undisciplined problems. In Interdisciplinary analysis and research (vol. 11, pp. 67–83). Lomond Publications, Mount Airy. Stern, S. (1989). Measuring the effect of disability on labor force participation. Journal of Human Resources, 24(3), 361–395. Van Noorden, R. (2015). Interdisciplinary research by the numbers. Nature, 525(7569), 306–307. Viseu, A. (2015). Integration of social science into research is crucial: Social scientists must be allowed a full, collaborative role if researchers are to understand and engage with issues that concern the public. Nature, 525(7569), 291–292. Weingart, P., & Stehr, N. (Eds.). (2000). Practising interdisciplinarity. Toronto: University of Toronto Press.

123