An effect of international students on the research productivity of American institutions

• In Event: Roundtable Session (Tuesday V)
  In Refereed Round-Table Session: Partnerships in research and knowledge production

Proposal

America is a leading country of internationalization of higher education. The US is the top destination of international students (Chow & Marcus, 2007; Lassegard 2016; Ren & Hagedorn, 2012; She & Wotherspoon, 2013; van der Wende, 2015). The financial contributions of international students reached $ 36.9 billion in 2016/17 (Choudaha, 2018). Internationalization has increased an importance of student mobility and research activities of universities in the US.
Having more international students can both encourage and impede the research activities of universities. International students have assumed a crucial role in advancing scientific research and technology in the US (Pandit, 2007). International students tend to be more productive (Chellaraj, Maskus, & Mattoo, 2008; Tran, 2016). They can enrich campus culture and deliver diverse perspectives for research (Ottaviano & Peri, 2006; Sparber, 2009). Moreover, because they pay higher tuition fees, institutions can earn more revenue and allocate more financial resources to research activities (Altbach & Knight, 2016; Burnett & Huisman, 2010; Chan & Dimmock, 2008; Childress, 2009; Doerpinghaus, Miller, & Pritt, 2018; Gopal, 2014; McFadden, Maahs-Fladung, & Mallett, 2012; Owens, Srivastava, & Feerasta, 2011; Scott, Safdar, Desai Trilokekar, & El Masri, 2015; Tremblay, 2005). Nonetheless, professors may not have enough time to conduct research when they have more numbers of students to teach and advise. Furthermore, universities may have to spend more money on operating additional services for international students (Ishikawa, 2011; Marginson & Rhoades, 2002). International students may not be necessarily beneficial to universities.
Therefore, the purpose of this paper is to investigate whether having more international students can directly predict the higher research productivity of an American institution or indirectly influence the research productivity by generating more revenue of an institution. This paper tries to investigate how the international students can predict one institution’s research productivity.
Previous studies have found that financial support (Adams & Clemmons, 2006; Defazio, Lockett, & Wright, 2009; Fedderke & Goldschmidt, 2015; Muscio, Quaglione, & Vallanti, 2013; Webber, 2012; White, James, Burke, & Allen, 2012; Wills, Ridley, & Mitev, 2013; Woldegiyorgis, Proctor, & de Wit, 2018; Zhang, Bao, & Sun, 2016), teaching load (Adams & Clemmons, 2006; Dundar & Lewis, 1998; Graves, Marchand, & Thompson, 1982; Mamiseishvili, 2010; Mamiseishvili & Rosser, 2010; Webber, 2012; White et al., 2012; Wills, Ridley, & Mitev, 2013), the number of professors (Dundar & Lewis, 1998; Zhang, Bao, & Sun, 2016), and the number of graduate students (Dundar & Lewis, 1998; Kwon, Kim, Park, Kim, & Jang, 2015; White et al., 2012; Zhang, Bao, & Sun, 2016) would predict the research productivity of an institution.
To meet the purpose, I conducted path analysis with a few variables discussed on the previous studies. I chose the number of domestic undergraduate and graduate students, the number of international undergraduate and graduate students, R&D expenditure, expenditure in library, and the number of faculty members as independent variables. I assumed that the number of students will have both a direct effect on the research productivity and an indirect effect on the research productivity via R&D expenditure. I also added research revenue as a control variable of R&D expenditure. All the data for the independent variables were extracted from the Integrated Postsecondary Education Data System (IPEDS). For human resources, I averaged the data of the past three years, 2014, 2015, and 2016. For financial resources, I accumulated the past three years’ data. This is because the publication of a research paper needs some time (Adams & Clemmons, 2006; Tie, 2012). However, I cannot distinguish master’s and doctoral students because the IPEDS did not specify the related data.
The dependent variable is the research productivity of an American institution. Because it is nearly impossible to accurately calculate the number of papers published by an institution (Zhang, Sun, & Bao, 2017), instead, I used the scores of publication (PUB) that each American institution obtained in the Academic Ranking of World Universities (ARWU) 2018. Lastly, I transformed the variables by taking log to meet the regression assumptions.
Initially, the ARWU 2018 listed 217 American institutions. However, I excluded four institutions that did not have data on the IPEDS. The correlation analysis result shows that all variables have statistically significant positive correlations except the correlation between the number of domestic graduate students and research revenue. Mostly, the correlations are moderate.
Then, I conducted path analysis. Checking the model fit, I could decide that the model was adequate as GFI and CFI are greater than .90.
I interpret standardized coefficients of the model. Domestic undergraduate students negatively predicted the number of publications (β = -.162, p < .05) but did not significantly predict R&D expenditure. Domestic graduate students did not statistically predict both R&D expenditure and the publications. International undergraduate students negatively predicted R&D expenditure (β = -.303, p < .05) and publication (β = -.256, p < .01). On the contrary, international graduate student positively predicted both R&D expenditure (β = .422, p < .001) and the number of publications (β = .386, p < .001). Library expenditure (β = .547, p < .001) and faculty members (β = .306, p < .001) positively predicted the number of research papers. Research revenue statistically significantly predicted R&D expenditure (β = .392, p < .001). R&D expenditure marginally predicted the publication (β = .078, p < .10).
The negative influence of undergraduate students on research looks obvious because professors will have higher teaching loads when they have more students to teach. International undergraduate students had a negative influence on R&D expenditure. This may be because institutions had to spend some money on services for them (Ishikawa, 2011; Marginson & Rhoades, 2002).
This result shows the importance of international graduate students for research. While domestic graduate students were not statistically significant, international students had a positive effect on the publications. Because R&D expenditure marginally predicted the publications, international graduate students rather directly affect the publications. As previous studies argued, this may be because they are talented and bring new perspectives and skills that enhance research productivity (Niebuhr, 2010; Tran, 2016).

Author

• Wan Yu, Penn State University