- Analytics: An overarching concept that is defined as data-driven decision making;
- Academic Analytics: A process for providing higher education institutions with the data necessary to support operational and financial decision-making;
- Learning Analytics: The use of analytics techniques to help target instructional, curricular and support resources to support the achievement of specific learning goals; and
- Predictive Analytics: An area of statistical analysis that deals with extracting information using various technologies to uncover relationships and patterns within large volumes of data that can be used to predict data and events.
The diversity of its use opens up analytics to professionals across higher education, from enrollment managers and academic administrators to front-line student success professionals like academic advisors. Despite its utility, though, this new and diverse field of practice can be perceived as uncomfortable or unreliable by practitioners who have yet to engage with data in this manner.
What Makes Analytics Difficult
Many of the challenges associated with the use of analytics are directly related to barriers caused by the improper organization and retrieval of data that is stored and used every day. While most colleges and universities collect voluminous amounts of student information, that data is not typically compiled strategically—instead departments and administrative offices focus on collecting what is standardized, convenient and accessible. Unfortunately, institutions are then stuck with a lot of information, but without the “right” information to investigate issues broadly or to design effective interventions.
Data collection issues may also originate from varying institutional terminology that impacts who collects the information and how it is collected. For example, when a college wants to track the success of student veterans, students may be classified using a self-reported metric from a survey, or by an enrollment management officer choosing to use the various categories of awarded federal VA benefits to identify which students are veterans. While both methods are valid, different sources and definitions for common terms can cause confusion and perceived inaccuracies.
Finally, some administrators mistakenly believe that simply providing access to technologies that deliver evaluations via spreadsheets, or which group students into categories using predictive modeling, is enough to drive changes in decision making. Oftentimes, end users are uncertain what the data is telling them or how to act on the data once it’s received.
Without knowing the underlying issue behind the evaluation or context of the data, many higher education professionals react to analytics with a simple, “That’s nice. What now?”
Effective Partnerships with the IT Industry
Because analytics can be challenging to develop, many colleges and universities seek applications designed by the numerous vendors within the higher education technology market. These products can be particularly beneficial when an institution is collecting large quantities of data without translating it into actionable information. Purchasing an out-of-the-box analytics product is another option for gaining perspective on institutional-level questions: Is this issue systematic for all enrolled students or an isolated problem affecting one sub-population? Data visualization resources are an excellent strategy for making analytics more accessible for end users with limited data literacy or limited time to actively engage with data. (**Pro-tip: Visualizations can sometimes entice the most cynical data skeptics into giving analytics a chance!)
After colleges and universities make the decision to purchase applications to generate analytics, it is vital to have intentional conversations with vendors about barriers to the effective use of analytics and develop strategies early in the implementation process to mitigate any potential risks. Remember to ask questions BEFORE committing to technology contracts in order to evaluate the options and resources available to support your institutional goals.
Vendors should be able to:
- Provide Transparency: Yes, predictive modeling is considered proprietary information; however, the vendor’s data scientists should be willing and able to meet with your institution’s data wonks to confirm the model is valid and reliable according to accepted data methodology.
- Define Terminology: Every vendor will have different language to describe their product and will operationalize data differently. Having this information provides context to the analytics (e.g., is persistence defined as being registered for at least one class for the next term or being registered for a full-time schedule?).
- Ensure Data Privacy: There must be clear methods to protect students against the misuse of information while still ensuring it is easily accessible to those who need it.
Postsecondary institutions must also acknowledge that purchasing technology does not change culture and should actively work to adjust processes to support data-informed practice. Although results from the models and analytics may shed light on where opportunities for improvement exist, administrators and staff are still often asked how and when to make those improvements. Analytics cannot solve the problems in a vacuum.
Finally, no matter how many dazzling donut charts and bubble plots a system can provide, an analytics tool only has value if it illuminates the needs of your students and empowers an institution’s ability to support student success.
– – – –
References
Van Barneveld, Angela & E Arnold, Kimberly & P Campbell, John. (2012). Analytics in Higher Education: Establishing a Common Language. Educause Learning Initiative. 1. Page 8. Retrieved from https://library.educause.edu/resources/2012/1/analytics-in-higher-education-establishing-a-common-language.
You Might Also Like
