Abstract
Postsecondary students increasingly enroll in online courses, which have the potential to further democratize higher education by expanding access for historically underserved populations. While a number of studies have investigated student outcomes in online courses, past data limitations have hindered robust examination of a potential mechanism underlying the decision to enroll in an online course: access to high speed broadband. With data from the National Broadband Map and IPEDS, I fit a number of Bayesian regression models to investigate the relationship between various measures of broadband access—download speed, upload speed, and the number of providers—and the number of students who take online courses at public colleges and universities with open admissions policies. Results show that increases in broadband speed at the lower end of the speed spectrum are positively associated with the number of students who take some of their courses online, but that the marginal gain diminishes as speeds increase. This finding suggests that there may be a minimum threshold of necessary broadband access, beyond which increases in speed become a less important factor in the take up of online coursework. Open admissions colleges seeking to improve access for local students through increased online course offerings should consider broadband access in the area, particularly if the targeted populations live in communities with low average broadband speeds.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
cf. the Open University and its founding as an autonomous institution.
Theoretical justifications notwithstanding, it is important to note that when weakly-informative priors are used in applied analyses, otherwise comparable Bayesian and frequentist models generally produce substantively similar estimates. Prior distributions are discussed later in the section.
A Bayesian posterior distribution is properly written as \(\begin{aligned} p(\theta \mid X)&= \frac{p(\theta )p(X\mid \theta )}{p(X)}, \end{aligned}\) which includes the evidence, p(X), in the denominator. Because data are assumed fixed in many Bayesian analyses, p(X) may be treated as a normalizing constant and dropped, yielding, \(p(\theta \mid X) \propto p(\theta )p(X\mid \theta )\).
When coding the model, I technically use improper priors. Priors are improper when \(\int p(\theta )\,d\theta \ne 1\), that is, the probabilities do not sum to one. All regression coefficients are drawn from a uniform distribution with support on \(\theta \in (-\infty , \infty )\) and all variances from a positive uniform distribution: \(\theta \in (0, \infty )\). Improper priors may combine with a likelihood function, however, to produce proper posterior distributions (Gelman et al. 2014).
In an alternative specification, I use the proportion of students as the outcome of interest. To properly model the proportion, which is bounded by [0,1], I use a beta likelihood function. The results for these models are qualitatively the same as those given by the log outcome/normal likelihood models, so I present the latter for ease of interpretation. More details about the beta likelihood specification as well results from the models are shown in Appendix 1.
IPEDS variable: EFDESOM.
Though subsequent years of student enrollment data have since been released, the National Broadband Map stopped being updated in June 2014. I have not incorporated the newest enrollment data for this reason.
Because the census block group population values are constant, it is not strictly necessary build an entire \(N\times K\) matrix, which is simply a vector of values of length K repeated N times, or compute the row sums and weights N times. For purposes of computation, however, it is easier to build a full matrix that can be easily combined with the other weighting matrix when computing the final average.
Combined weighted broadband equation: \(\begin{aligned} wbroadband_{s} = \sum _{c = 1}^{C}\frac{ \Big (\frac{pop_{c}}{\sum _{c = 1}^C pop_{c}}\Big ) \Big (\frac{id_{sc}}{\sum _{c = 1}^C id_{sc}}\Big ) \cdot broadband_{c}}{\sum _{c = 1}^{C} \Big (\frac{pop_{c}}{\sum _{c = 1}^C pop_{c}}\Big ) \Big (\frac{id_{sc}}{\sum _{c = 1}^C id_{sc}}\Big )} \end{aligned}\)
All models were estimated using (CmdStan, version 2.18.0), the command line version of Stan’s No-U-Turn Sampler (NUTS), a variant of the Hamiltonian Monte Carlo sampler that may more efficiently explore the parameter space. To reduce the amount of lagged auto-correlation between successive draws in the chains, improving convergence and the number of effective samples, all models were estimated using centered data and a QR reparameterization (Stan Development Team 2018; Lunn et al. 2013).
All reported 95% credible intervals represent the middle 95% of the posterior distribution, meaning that the lower and upper bound values are the 2.5% and 97.5% quantile values, respectively.
Throughout the rest of the chapter, I will generally use the Bayesian point estimates when referring to the mean of parameter’s posterior distribution. They should be understood, therefore, in their proper context as useful summaries of full probability distributions.
The margins in this section are computed using the full posterior distribution of the main and quadratic parameters. They are therefore slightly different than what is computed using only the Bayesian point estimates presented in the results tables.
If the mean value presented in the table represented a frequentist point estimate, one could not reject the null that \(\beta = 0\) under a two-tailed test of significance at conventional levels of significance (\(\alpha = 0.05\)). This is not to say, however, that it would not be jointly significant with its quadratic term.
These were computed from the data by taking the fractional value of the minimum upload speed over the minimum download speed, the same fractional value of the relative maximum speeds, and building a sequence of 11 values (1 to 11 on the NBM scale) that range between them.
While the focus of this paper has been on public institutions with open admissions policies, I also ran all models using the full census of colleges and universities during this period who reported having any students take some courses online as part of a sensitivity analysis. All models were the same, but with the inclusion of indicators for being private-non-profit, private-for-profit, or having an open admissions policy. Findings were qualitatively similar to those discussed in the paper. Tables showing these results are included in Appendix 2.
References
Allen, I. E., & Seaman, J. (2011). Going the distance. Tech. rep. Babson Survey Research Group.
Allen, I. E., & Seaman, J. (2016). Online report card: Tracking online education in the United States. Tech. rep. Babson Survey Research Group
Becker, G. S. (2009). Human capital: A theoretical and empirical analysis, with special reference to education (3rd ed.). Chicago: University of Chicago Press.
Bettinger, E. P. (2017). Virtual classrooms: How online college courses affect student success. American Economic Review, 107(9), 2855–75.
Bowen, W. G. (2013). Higher education in the digital age. Princeton, NJ: Princeton University Press.
Bowen, W. G., Chingos, M. M., Lack, K. A., & Nygren, T. I. (2014). Interactive learning online at public universities: Evidence from a six-campus randomized trial. Journal of Policy Analysis and Management, 33(1), 94–111
Brooks, S. (2011). Handbook of Markov Chain Monte Carlo. Handbooks of modern statistical methods. Boca Raton: Chapman and Hall.
Brown, R. H., Barram, D. J., & Irving, L. (1995). Falling through the net: A survey of the ‘Have Nots’ in rural and urban America. Tech. rep. Washington, D.C.: National Telecommunications and Information Administration, U.S. Department of Commerce.
Brown, T., Chawla, M., Curtis, R., Dixit, R., Goodheart, R., Kendall, T., et al. (2010). The broadband availability gap. OBI Tech. Paper 1. Federal Communications Commission.
Bureau of Labor Statistics. (2012). Unemployment Rates for States, 2012 Annual Averages. Online table.
Bureau of Labor Statistics. (2013). Unemployment Rates for States, 2013 Annual Averages. Online table.
Bureau of Labor Statistics. (2014). Unemployment Rates for States, 2014 Annual Averages. Online table.
Carlin, B. P., & Louis, Thomas A. (2009). Bayesian methods for data analysis (3rd ed.). Boca Raton: CRC Press.
Carlson, A., Ott, C., Armstrong, J., Zaback, K., & Auer, G. (2015). State Higher Education Finance FY 2014. Tech. rep. Boulder, CO: State Higher Education Executive Officers Association.
Cejda, B. D. (2007). Distance education in rural community colleges. Community College Journal of Research and Practice, 31(4), 291–303.
Cohen, A. M., & Brawer, F. B. (2003). The American Community College (4th ed.). Hoboken: Wiley.
Copps, M. J. (2009). Bringing broadband to rural America: Report on Rural Broadband Strategy. Federal Communications Commission.
Cotten, S. R., & Jelenewicz, S. M. (2006). A disappearing digital divide among college students? Peeling away the layers of the digital divide. Social Science Computer Review, 24(4), 497–506.
Cox, R. D. (2006). Virtual access. In T. Bailey & V. Smith Morest (Eds.), Defending the Community College equity agenda (pp. 110–131). Baltimore: Johns Hopkins University Press. Chap. 5,
Crookston, A., & Hooks, G. (2012). Community colleges, budget cuts, and jobs: The impact of community colleges on employment growth in rural U.S. counties, 1976–2004. Sociology of Education, 85(4), 350–372.
Czernich, N. (2011). Broadband infrastructure and economic growth. The Economic Journal, 121(552), 505–532.
Deming, D. J. (2015). Can online learning bend the higher education cost curve? American Economic Review, 105(5), 496–501.
Eagan, K., Stolzenberg, E. B., Ramirez, J. J., Aragon, M. C., Suchard, M. R., & Hurtado, S. (2014). The American Freshman: National Norms Fall 2014. Tech. rep. Los Angeles, CA: Higher Education Research Institute at UCLA.
Federal Communications Commission. (2009). Connecting America: The National Broadband Plan. Federal Communications Commission.
Federal Communications Commission. (2014). Measuring Broadband America Fixed Broadband Report: A Report on Consumer Fixed Broadband Performance in the U.S. Tech. rep. Washington, D.C.: Federal Communications Commission; Office of Engineering.
Federal Communications Commission. (2015). 2015 Broadband Progress Report. FCC 14-113. Washington, D.C.: Federal Communications Commission.
Federal Communications Commission. (2016a). 2016 Broadband Progress Report. FCC 16-6. Washington, D.C.: Federal Communications Commission.
Federal Communications Commission. (2016b). Measuring broadband America fixed broadband report: A report on consumer fixed broadband performance in the United States. Tech. rep. Washington, D.C.: Federal Communications Commission; Office of Engineering.
Figlio, D. N., Rush, M., & Yin, L. (2010). Is it Live or is it Internet? Experimental Estimates of the Effects of Online Instruction on Student Learning. Working Paper 16089. National Bureau of Economic Research.
Fishman, R. (2015). Deciding to go to college. 2015 College Decisions Survey: Part 1. Washington, D.C.: New America Education Policy.
Gelman, A. (2014). Bayesian data analysis (3rd ed.). Boca Raton: CRC Press.
Gill, J., & Walker, L. D. (2005). Elicited priors for Bayesian model specifications in political science research. The Journal of Politics, 67(3), 841–872.
Goldrick-Rab, S. (2010). Challenges and opportunities for improving community college student success. Review of Educational Research, 80(3), 437–469.
Greene, W. H. (2012). Econometric Analysis (7th ed.). Boston, MA: Prentice Hall.
Grubesic, T. H. (2008a). The spatial distribution of broadband providers in the United States: 1999–2004. Telecommunications Policy, 32(3–4), 212–233.
Grubesic, T. H. (2008b). Spatial data constraints: Implications for measuring broadband. Telecommunications Policy, 32(7), 490–502.
Grubesic, T. H. (2008c). Zip codes and spatial analysis: Problems and prospects. Socio-Economic Planning Sciences, 42(2), 129–149.
Grubesic, T. H. (2012a). The wireless abyss: Deconstructing the U.S. National Broadband Map. In: Government Information Quarterly 29.4, pp. 532 -542.
Grubesic, T. H. (2012b). The U.S. National broadband map: Data limitations and implications. Telecommunications Policy, 36(2), 113–126.
Grubesic, T. H., & Murray, A. T. (2002). Constructing the divide: Spatial disparities in broadband access. Papers in Regional Science, 81(2), 197–221.
Grubesic, T. H., & Murray, A. T. (2004). Waiting for broadband: Local competition and the spatial distribution of advanced telecommunication services in the United States. Growth and Change, 35(2), 139–165.
Hanigan, I., Hall, G., & Dear, K. B. G. (2006). A comparison of methods for calculating population exposure estimates of daily weather for health research. International Journal of Health Geographics, 5(1), 38.
Hart, C. M. D., Friedmann, E., & Hill, M. (2018). Online course-taking and student outcomes in California Community Colleges. Education Finance and Policy, 13(1), 42–71.
Hoffman, D. L., & Novak, T. P. (1998). Bridging the racial divide on the Internet. Science, 280, 390–391.
Horrigan, J. B. (2010). Broadband adoption and use in America. OBI Working Paper Series 1. FCC.
Huntington-Klein, N., Cowan, J., & Goldhaber, D. (2017). Selection into online community college courses and their effects on persistence. Research in Higher Education, 58(3), 244–269.
Hurst, M. A. (2010). Residential Broadband Access for Students at Walters State Community College. Dissertation. East Tennessee State University.
Irving, L., Klegar-Levy, K., Everette, D. W., Reynolds, T., & Lader, W. (1999). Falling through the net: Defining the digital divide. Tech. rep. Washington, D.C.: National Telecommunications & Information Administration, U.S. Department of Commerce.
Jackman, S. (2009). Bayesian analysis for the social sciences. Wiley series in probability and statistics. London: Wiley.
Jaggars, S. S. (2014). Choosing between online and face-to-face courses: Community college student voices. American Journal of Distance Education, 28(1), 27–38.
Johnson, J. L. (2003). Distance education: The complete guide to design, delivery, and improvement. New York: Teachers College Press.
Johnstone, D. (2006). The rise of Western Governors University, 1996–2006. Salt Lake City: Western Governors University.
Jones, S. (2009). U.S. college students’ Internet use: Race, gender and digital divides. Journal of Computer-Mediated Communication, 14(2), 244–264.
Joyce, T. J., Crockett, S., Jaeger, D. A., Altindag, O., & O’Connell, S. D. (2014). Does classroom time matter? A randomized field experiment of hybrid and traditional lecture formats in economics. Working Paper w20006. Cambridge, MA: National Bureau of Economic Research.
Keegan, D. (1996). Foundations of distance education (3rd ed.). New York, NY: Routledge.
Kinser, K. (1999). The Origin, Development, and Implications of the Western Governors University. Doctoral dissertation. Columbia University.
Loss, C. P. (2012). Between citizens and the state: The politics of American higher education in the 20th century. Princeton: Princeton University Press.
Lunn, D. (2013). The BUGS book: A practical introduction to Bayesian analysis. Boca Raton: CRC Press.
Manski, C. F., & Wise, David A. (1983). College choice in America. Cambridge, MA: Harvard University Press.
Mattern, K., & Wyatt, J. N. (2009). Student choice of college: How far do students go for an education? Journal of College Admission, 203, 18–29.
McConnaughey, J. W., & Lader, W. (1998). Falling through the net: New data on the digital divide. Tech. rep. Washington, D.C.: National Telecommunications & Information Administration, U.S. Department of Commerce.
Moloney, J. F., & Oakley, B. (2010). Scaling online education: Increasing access to higher education. Journal of Asynchronous Learning Networks, 14(1), 55–70.
Moore, M. G., & Kearsley, G. (2011). Distance education: A systems view of online learning. Belmont: Cengage Learning.
National Center for Education Statistics Department of Education. (2015). Integrated Postsecondary Education Data System (IPEDS). Data file.
National Center for Education Statistics Department of Education. IPEDS Glossary. (Visited on 2015).
National Telecommunications and Information Administration. National Broadband Map. (Visited on 2011).
Ortagus, J. C. (2018). National evidence of the impact of first-year online enrollment on postsecondary students’ long-term academic outcomes. In Research in Higher Education.
Oyana, T. J. (2011). Exploring geographic disparities in broadband access and use in rural southern Illinois: Who’s being left behind? Government Information Quarterly, 28(2), 252–261.
Pappano, L. (2012). The Year of the MOOC. In The New York Times.
Prieger, J. E., & Wei-Min, H. (2008). The broadband digital divide and the nexus of race, competition, and quality. Information Economics and Policy, 20(2), 150–167.
Radford, A. W., & Weko, T. (2011). Learning at a Distance: Undergraduate Enrollment in Distance Education Courses and Degree Programs. NCES 2012-154. Washington, D.C.: U.S. Department of Education, National Center for Education Statistics, Institute of Education Sciences.
Rosenboom, V., & Blagg, K. (2018). Disconnected from higher education. Tech. rep. Urban Institute:
Selingo, J. J. (2013). College (Un)Bound. New York: Houghton Mifflin Harcourt Publishing Company.
Shepard, D. (1968). A two-dimensional interpolation function for irregularly-spaced data. Proceedings of the 1968 23rd ACM national conference (pp. 517–524). ACM.
Snyder, T. D., de Brey, C., & Dillow, S. A. (2016). Digest of Education Statistics: 2015. NCES 2016-014. Washington, D.C.: National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education.
Snyder, T. D., & Dillow, S. A. (2015). Digest of Education Statistics: 2013. NCES 2015-011. Washington, D.C.: National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education.
Southern Region Educational Board. (2013). State Data Exchange: 2011–2012 Indicators Report. Southern Regional Education Board.
Southern Region Educational Board. (2015). State Data Exchange: 2013–2014 Indicators Report. Southern Regional Education Board.
Stan Development Team. (2018). Stan modeling language users guide and reference manual. Version 2.18.0.
Tandberg, D. A. (2010). Politics, interest groups and state funding of public higher education. Research in Higher Education, 51(5), 416–450.
The Open University. The OU story. (Visited on 2015).
Thelin, J. R. (2011). A history of American higher education (2nd ed.). Baltimore, MD: The Johns Hopkins University Press.
Turner, S. E. (2004). Going to college and finishing college: Explaining different educational outcomes. In M. Caroline (Ed.), College choices: The economics of where to go, when to go, and how to pay for it (pp. 13–62). Chicago: University of Chicago Press.
United States Department of Agriculture Economic Research Service. (2013). 2013 Rural–Urban Continuum Codes. Online data set.
U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics. (2012). National Postsecondary Student Aid Study, 2012 (NPSAS:12). Online data set.
Western, B., & Jackman, S. (1994). Bayesian inference for comparative research. The American Political Science Review, 88(2), 412–423.
Xu, D., & Jaggars, S. S. (2011). The effectiveness of distance education across Virginia’s Community Colleges: Evidence from introductory college-level math and English courses. Educational Evaluation and Policy Analysis, 33(3), 360–377.
Xu, D., & Jaggars, S. S. (2013). The impact of online learning on students’ course outcomes: Evidence from a large community and technical college system. Economics of Education Review, 37, 46–57.
Acknowledgements
This research was supported in part by the Bonsal Applied Education Research Award. I thank two anonymous reviewers as well as Angela Boatman, Josh Clinton, Stephen DesJardins, Will Doyle, Carolyn Heinrich and participants at the 2015 ASHE annual conference for their helpful comments and suggestions. This article is better for their help. All findings and conclusions remain my own.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix 1: Alternative Specifications
In an alternative specification, I fit equations using the proportion of students who took some online courses as the outcome. To model this outcome, I use a beta likelihood function with dispersion parameter that accurately accounts for the [0,1] bounds of the dependent variable (Gelman et al. 2014):
As with the primarily analysis models, I fit single level as well as multilevel models in which the intercepts and parameters on broadband measures are allowed to vary at the state level. Results from these models show the marginal effect of broadband on the logged-odds of a percentage increase in the number of students who took some online courses. The results from these models are qualitatively the same as those reported in the main text of the chapter (Tables 5, 6).
Appendix 2: Sensitivity Analyses
See Tables 7, 8, 9, 10, 11, and 12.
Rights and permissions
About this article
Cite this article
Skinner, B.T. Making the Connection: Broadband Access and Online Course Enrollment at Public Open Admissions Institutions. Res High Educ 60, 960–999 (2019). https://doi.org/10.1007/s11162-018-9539-6
Received:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11162-018-9539-6