The benefits of learning analytics are manifold. Learners can get real-time information on their progress and learning gaps; instructional designers and facilitators can be better informed regarding how learning activities are received by individuals and groups of learners in order to adjust learning programs accordingly. Additionally, managers are able to glean insights on how learning potentially supports and translates to on-the-job performance.
The underlying assumption of learning analytics is that we will use the data collected to gain insights on the learning activities and learner behaviors, interpret the data, and provide interventions and predictions. Typically, much of the data comes from the learning management system (LMS), but also through informal learning channels, face-to-face session attendance records and evaluations, access to other content repositories, and even mobile phone usage.
However, when it comes to implementing learning analytics, we are faced with many data problems. The following factors outline a few of the problems surrounding learning analytics and potential solutions:
The main problem with the use of data in learning: We need to have them. Many learning and development (L&D) departments still rely on paper-based records, some don’t get past the collection of smile sheets for training evaluations and some are not in the habit of collecting any learning data. Often, data are out of date, stored across different databases and formats, and are in general quite messy. It is hard enough just getting the data; we need to take the first step in collecting data in a format that is structured, cohesive and standardized. There are tools and data science professionals to help you with data cleansing using statistical software.
Different Types of Data
We need to get past the Shareable Content Object Reference Model (SCORM) in terms of the types of data to collect. SCORM measures learning activity completion, but completion doesn’t prove that learning has taken place. To identify if the desired learning outcomes are being achieved, learning professionals need to start with the end in mind when designing learning activities and programs. What kind of interactions and assessments can you design that allow you to measure learning meaningfully? One solution is to design formative self-knowledge checks throughout an online course to spot whether learners have grasped the concepts or if additional supports are needed in real time rather than waiting until the end of the course.
I Have Data and I Need Insights
There is a fundamental flaw in approaching learning analytics this way. We should be asking: What learning problem and, ultimately, performance problem am I trying to solve that data can help provide insights to? For example, you are trying to solve the problem of customers’ complaints about the poor service provided by your call center staff. You might hypothesize that, by providing a customer service training simulation experience for the staff, the number of complaints from customers will go down by a certain percent. Does the data reflect that? If not, you must figure out the root cause of what is not working.
Data is Not Knowledge
Knowledge, at minimum, involves a practical understanding of subjects, an acquisition of skills, and a demonstration of competencies or mastery level applications. Data collected from various learning sources and systems haven’t undergone thorough processing and are meaningless. While it is important to collect diverse sets of data, it is from the interpretation and analysis of the data that we can infer how learners learn and the conditions needed for successful learning. Furthermore, once you’ve gained these learning insights, the question becomes: How do you convert insights into actions?
All in all, learning analytics represent a real opportunity for learning professionals to provide evidence-based learning interventions, enhance learning experiences and foster better performance support. Understanding and applying data strategically will go a long way in the successful implementation of learning analytics at your workplace.