The posting below looks at some key elements to pay attention to in taking advantage of the massive amounts of information released by the analytics revolution in higher education. It is from Chapter 1: The Analytics Revolution in Higher Education, by Jonathan S. Gagliardi, in the book, Big Data, Organizational Learning, and Student Success, edited by Jonathan S. Gagliardi, Amelia Parnell, and Julia Carpenter-Hubin. Published by Stylus Publishing, LLC 22883 Quicksilver Drive Sterling, Virginia 20166-2102. https://sty.presswarehouse.com/books/features.aspx Copyright © 2018 by Stylus Publishing, LLC. All rights reserved. Reprinted with permission.
UP NEXT: Do English-Only Policies Foster or Damage Inclusivity in Science?
---------- 2,185 words ----------
The Analytics Revolution in Higher Education
We are living through an analytics revolution. More data was created over the last two years than in the entire previous history of humankind1 (IBM, 2016; Marr, 2016). Those numbers will be dwarfed in the coming years as the digital universe grows to include over 50 billion devices and over 180 billion zettabytes of data by the year 2025 (IDC, 2014; Marr, 2016). For example, transmitting all that data through a broadband Internet connection would take over 450 million years, according to a recent article in the Economist (“Data Is Giving Rise,” 2017). By the turn of the decade, approximately 1.7 megabytes of new information will be created every second for each person on Earth (Evans, 2011; Gantz & Reinsel, 2012; Marr, 2016).
Still, raw data is of limited utility. To extract value, it needs to be mined, refined, integrated, and analyzed for insight. When responsibly and effectively used, the insight drawn from data propels progress and innovation. The potential benefits are too many to ignore for people, firms, and governments, many of which have taken measures to accelerate their analytics maturity. Successfully doing so has become big business in recent years. In 2015, revenues from analytics software reached $122 billion, and that number is expected to rise to $187 billion in 2019. The market for predictive and prescriptive analyses is skyrocketing (Columbus, 2016). More firms are recognizing the value of putting data to work to promote organizational transformation. So powerful is this analytics revolution that in 2013 the term big data was added to the Oxford English Dictionary. A fad this is not; the analytics revolution is here to stay.
While this analytics revolution has been unfolding, American colleges and universities have been facing choppy water. The U.S. system of higher education, highly regarded by many, has shown signs of dysfunction in recent years. In fact, America is no longer the most highly educated nation in the world (OECD, 2016). This can be attributed, in part, to the convergence of stagnant graduation rates, persistent equity gaps, ballooning tuition, mounting student debt, and other countries getting serious about postsecondary education. These factors, coupled with social, political, and economic changes, have weakened the public trust in higher education, eroded state and federal financial support for the pursuit of a college credential, and created an affordability problem. Sensing a crisis, stakeholders from government, the private sector, and civic society have renewed their calls for higher education reform by demanding transparency and accountability and creating lofty educational attainment goals known collectively as the completion agenda. Still, improvement has been a slow and difficult process.
Colleges and universities have tried to reassess and reconfigure their business models in hopes of better serving students, communities, and economies in response to this growing crisis (Soares, Steele, & Wayt, 2016). These efforts to improve student outcomes while driving down costs have primarily focused on the large-scale adoption of programs, practices, and services designed to optimize remediation, shorten time to degree, reduce excess credits, and streamline credit transfer, all while enhancing teaching, learning, and advising in a cost-effective manner (Complete College America, 2017). Doing so can be hard because institutions are trying to standardize output with fewer resources and increasingly varied input, but it is not impossible. Data analytics are at the heart of gathering the evidence and insights needed to accomplish the transformational changes demanded by the current climate. In recent years, robust data analytics have been shown to be a key ingredient to strategic innovation.
Even though the higher education community has turned the corner and embraced the analytics revolution, a multitude of barriers stand in the way of any given campus doing so. While data and analytics tools are plentiful, the reality is that most institutions are not yet able to use them optimally, for a host of reasons. These can include: insufficient or misaligned resources, endless information demands, disjointed or rigid infrastructure, mismatches in skills and expertise, and the absence of data-enabled executives (Gagliardi & Turk, 2017). The presence of any of these challenges can be formidable enough to undermine the development of an analytics culture, and a large share of campuses experience a handful of them. So, how can colleges and universities overcome these barriers to harness the power of data analytics?
This volume seeks to help college and university leaders2 harness the analytics revolution in ways that promote student success, organizational vitality, and innovation. The rest of this chapter will provide the reader with an overview of some of the challenges and opportunities facing colleges and universities in their quest to harness the analytics revolution. It will then home in on how institutional research (IR) is being reshaped by this fundamental shift in how data are being used in higher education. It is important to note that this volume does not focus on methods, analyses, or visualizations; rather, it focuses on identifying the keys to developing a dynamic, campus-wide analytics culture and function centered around IR.
Harnessing the Analytics Revolution
If effectively harnessed, the massive amount of information born out of the analytics revolution can allow institutions to better understand student needs; enhance the quality of teaching, learning, and advising; drive down costs; and predict and avoid risks (Cai & Zhu, 2015; Denley, 2014). Despite this, the sheer volume of data and the multitude of tools for analysis are inconsequential to colleges and universities unless the conditions exist for their effective use. In fact, while artificial intelligence and machine learning grab headlines, most colleges lack that level of analytics sophistication, and they do not need it. Many institutions would benefit from a solid foundation of data that is based on accuracy, timeliness, relevancy, integration, and security.
As the sheer volume of available data increases so do pressures to use it, making it important to develop procedures to ensure that it is of quality and usable in contextualized ways. There are multiple steps in acquiring, processing, and analyzing data. These include data discovery, extraction, reformatting, uploading, normalizing, enriching, comparing, presenting, and integrating workflow (Wheeler, 2017). These procedures help to ensure that new insights gleaned from data analytics are trustworthy ones.
Data and insight need to be delivered in timely and accessible ways, otherwise their usefulness may be lost regardless of their accuracy. This is especially true for colleges and universities that are seeking out real-time solutions to the challenges facing students. The longer it takes to acquire, process, and analyze data related to each element of the student lifecycle, the less likely it is that insights can be used for predicting risks and prescribing solutions for students.
Translating accurate and timely data into programs or services that support students and decision makers is often aspired to but seldom achieved. In some ways there is so much data it becomes difficult to sort the good from the bad. With so much data to sift through, the matter of identifying the right analytics tools and infrastructure becomes more important. As the need for accurate real-time insight increases, analytics leaders must be prepared to deliver insights, products, and services that matter to the end user.
Decision makers want access to insight in near-real time, which means that the steps of acquiring, processing, and analyzing data need to happen quickly. A major barrier to delivering accurate, timely, and relevant insight has been a lack of integration. The difficulties in sewing together data from disparate sources originate from a host of challenges, including differences in storage, definition, structure (or lack thereof), and intended use (Gagliardi & Wellman, 2015). At a time when unstructured data, which can be incredibly rich, account for 90% of enterprise data, determining ways to extract value often hinges on connecting it with other data sources (Vijayan, 2015). This makes effective integration an even more important step toward creating dynamic data and insight.
Indeed, it can be difficult to create data that are simultaneously accurate, timely, and integrated while also making them accessible to decision makers in real-time. On top of that, data need to be protected and to be used ethically. Policies and best practices surrounding data privacy and security, intellectual property, and ethical practices all warrant careful attention (Ekowo & Palmer, 2017). According to Sun (2014), analytics functions should work to adopt practice standards, adhere to best practices to maintain privacy and security, and create ethics review boards to mitigate the many risks associated with the analytics revolution, big data, and predictive analytics. Once data are accurate, timely, relevant, integrated, and secure, institutions can focus on building out the infrastructure and culture necessary to become more data-informed.
The right infrastructure is needed to acquire, process, and analyze data from diverse sources in relevant ways that are secure. In the Higher Education Data Warehousing Forum’s3 most recent survey of top issues facing its members,4 over half (57%) of respondents chose data governance as their top issue. Six of the top 12 categories were related to technology. These included data quality (45%), metadata and data definitions (42%), predictive analytics (35%), data visualization (34%), integration (33%), and self-service (30%) (Childers, 2017).
Investing in quality data, insight, and the underlying infrastructure requires that campuses reorient their cultures toward a collaborative model of data-informed decision-making. Without a culture of analytics, efforts to embed analytics can generate concerns around diluting quality, eliminating choice, tracking students, cutting programs and jobs, and the loss of institutional identity. Leadership must champion the use of data and be intentional about tying data analytics into a future vision focused on student success and institutional sustainability. The rewards are many for campuses willing to invest in infrastructure and culture, which are preconditions for safely and effectively harnessing the analytics revolution.
Improving Student Success
The major reward of better leveraging the analytics revolution is student success. For example, Austin Peay State University used a grade-prediction model to place students in courses that offered them the highest likelihood of success. The program, called Degree Compass, proved effective. Over 90% of students who took the recommended courses earned an A or a B. Moreover, the grades of students after Degree Compass was introduced were five standard deviations higher than those of students prior to its implementation. It was eventually scaled across the Tennessee Board of Regents. Since its initial deployment in 2011, similar models have been adopted at institutions across the country (Denley, 2014; Gagliardi, 2015).
In another instance, Georgia State University (GSU) increased its graduation rates from 32% in 2003 to 54% in 2015; graduation rate gaps for low-income, underrepresented, and first-generation students were also closed (Georgia State University, 2016). Between 2009-2010 and 2015-2016, the credit hours at completion for bachelor students decreased from 140 to 133. One contributing factor to GSU’s success was the analysis of 10 years of student financial data and approximately 140,000 GSU student records, which eventually became the baseline for the development of their predictive analytics platform (Georgia State University, 2016).
Each of these universities methodically built out their analytics capacity. Eventually they were able to use predictive analytics and principles of behavioral economics and choice architecture to better guide students along their education pathway with great success (Denley, 2014; Thaler & Sunstein, 2008; Wildavsky, 2014). The potential exists for all institutions to do so, as long as they know where to start, and IR is the ideal partner for figuring it out.
Opportunities for Institutional Research
IR has a leadership role to play in ensuring that campuses realize the full potential of the analytics revolution while mitigating the associated risks. This includes satisfying the growing sea of demands for insights that improve student outcomes and institutional productivity. The function is well positioned to guide the creation of knowledge, advise in policy and strategy, and drive innovation through a data-informed lens. This is because IR functions possess a unique combination of institutional memory and domain expertise, in addition to a rich understanding of the opportunities and challenges facing students and institutions (Cheslock, Hughes, & Umbricht, 2014). These characteristics make IR unique in its capacity to create relevant organizational intelligence that informs action.
In that vein, leaders of IR are exploring how to acquire, process, and analyze data in ways that lead to better decisions. They are also using the analytics revolution to create new kinds of visualization products, real-time business intelligence, and self-service platforms that appeal to leaders and power users alike. They are opening the door to more sophisticated analyses, better insight, and a more collaborative and nimble model of analytics led by IR.
In recognition of this emerging model, the Association for Institutional Research published a statement of aspirational practices for IR leaders and practitioners. It emphasizes a hybrid model of analytics (Swing & Ross, 2016). In this model, IR is charged with
· nurturing the profusion of IR across the institution,
· teaching data analytics best practices,
· facilitating campus-wide data-informed decision-making,
· providing professional development in data analytics, and
· focusing on student-centered responses.
By embracing these roles, IR leaders can modernize and enhance their capacity to guide institutions through an intentional process of transformational change (Swing & Ross, 2016). To do that, there are several challenges that must be overcome.
1. According to IBM, 2.5 quintillion bytes of data were created between 2014 and 2016.
2. These include senior executive roles such as presidents, provosts, chief institutional researchers, chief information officers, and other senior executives.
3. The Higher Education Data Warehousing Forum (HEDW) is a network of higher education colleagues dedicated to promoting the sharing of knowledge and best practices regarding knowledge management in colleges and universities, including building data warehouses, developing institutional reporting strategies, and providing decision support.
4. The HEDW’s members include technical developers and administrators of data access and reporting systems, data custodians, institutional researchers, and consumers of data representing a variety of internal university audiences.
Cai, L., & Zhu, Y. (2015). The challenges of data quality and data quality assessment in the big data era. Data Science Journal, 14, 2. DOI: http://doi.org/10.5334/dsj-2015-002
Cheslock, J., Hughes, R.P., & Umbricht, M. (2014). The opportunities, challenges, and strategies associated with the use of operations-oriented (big) data to support decision making within universities. In J. E. Land (Ed.), Building a smarter university: Big data, innovation, and analytics (pp. 211-238). Albany, NY: SUNY Press.
Childers, H. (2017, November 30). 2017 HEDW Survey of top 10 issues [Web log post]. Retrieved from http://hedw.org/2017-hedw-survey-of-top-10-issues/
Columbus, L. (2016, August 20). Roundup of analytics, big data & BI forecasts and market estimates, 2016. Forbes. Retrieved from https://www.forbes.com/sites/louiscolumbus/2016/08/20/roundup-of-analytics-big-data-bi-forecasts-and-market-estimates-2016/#79e616d26f21
Complete College America. (2017). The game changers. Retrieved from http://completecollege.org/the-game-changers/
Denley, T. (2014). How predictive analytics and choice architecture can improve student success. Research & Practice in Assessment, 9(2), 61-69.
Ekowo, M., & Palmer, I. (2017, March 6). Predictive analytics in higher education. New America. Retrieved from https://www.newamerica.org/education-policy/policy-papers/predictive-analytics-higher-education/#introduction
Evans, D. (2011). The Internet of things: How the next evolution of the Internet is changing everything [White paper]. Retrieved from http://www.cisco.com/c/dam/en_us/about/ac79/docs/innov/IoT_IBSG_0411FINAL.pdf
Gagliardi, J. S. (2015). From perpetuation to innovation: Breaking through barriers to change in higher education. In J. E. Land (Ed.), Higher education reconsidered: Executing change to drive collective impact (pp. 61-96). Albany, NY: SUNY Press.
Gagliardi, J. S., & Turk, J. M. (2017). The data enabled executive. Washington DC: American Council on Education.
Gantz, J., & Reinsel, D. (2012, December). The digital universe in 2020: Big data, bigger digital shadows, and biggest growth in the far east [Video file]. Retrieved from https://www.emc.com/leadership/digital-universe/2012iview/analyst-perspective-john-gantz-david-reinsel.htm
Georgia State University. (2016). 2016 status report. Retrieved from http://success.gsu.edu/files/2017/01/Georgia-State-University-2016-Complete-College-Report-with-Appendix-10-26-16.pdf
IBM. (2016). Big data. Retrieved from https://www.ibm.com/analytics/us/en/technology/big-data/
IDC. (2014). The digital universe of opportunities. Retrieved from https://www.emc.com/collateral/analyst-reports/idc-digital-universe-2014.pdf
Marr, B. (2016, November 1). 20 mind-boggling facts every business leader must reflect on now. Forbes. Retrieved from https://www.forbes.com/sites/bernardmarr/2016/11/01/20-mind-boggling-facts-every-business-leader-must-reflect-on-now/#35486aa20dcd
OECD. (2016). Education at a glance 2016: OECD indicators. Paris, France: OECD Publishing. http://dx.doi.org/10.1787/eag-2016-en
Soares, L., Steele, P., & Wayt, L. (2016). Evolving higher education business models: Leading with data to deliver results. Washington DC: American Council on Education.
Sun, J. C. (2014). Legal issues associated with big data in higher education: Ethical considerations and cautionary tales. In J. E. Lane (Ed.), Building a smarter university: Big data, innovation, and analytics (pp. 27-56). Albany, NY: SUNY Press.
Swing, R. L., & Ross, L. E. (2016). Statement of aspirational practice for institutional research. Association for Institutional Research. Retrieved from http://www.air-web.org/aspirationalstatement.
Vijayan, J. (2015, June 25). Solving the unstructured data challenge. CIO. Retrieved from http://www.cio.com/article/2941015/big-data/solving-the-unstructured-data-challenge.html
Wheeler, B. (2017, March 13). Who is doing our data laundry? EDUCAUSE Review. Retrieved from http://er.educause.edu/articles/2017/3/who-is-doing-our-data-laundry
Wildavsky, B. (2014). Nudge nation: A new way to use data to prod students into and through college. In J. E. Lane (Ed.), Building a smarter university: Big data, innovation, and analytics (pp. 143-158). Albany, NY: SUNY Press.