Analytics starts with measurement or the mere act of tracking things and recording values to tell learners what happened. Measurement does not need complicated math or statistics, but one must start by gathering data.
FREMONT, CA: Learning analytics is the measurement, compilation, analysis, and reporting of data about apprentices, learning programs and experiences for the purpose of understanding and enhancing learning and its effect on an organization’s performance.
The Levels of Learning Analytics
There are four levels of analytics: measurement, evaluation, advanced evaluation, and predictive and prescriptive analytics. Even though each of these levels is suita
bly referred to as analytics, they mean vastly different things in terms of complexity and power.
Analytics starts with measurement or the mere act of tracking things and recording values to tell learners what happened. Measurement does not need complicated math or statistics, but one must start by gathering data. Otherwise, it is impossible to do any analytics.
Upon capturing the data, it is time to start assessing it and evaluate whether the data means something good or bad. At this level, one applies high-school level math—means, modes, averages, and basic statistics—to aggregate the data and create benchmarks. In current practice, most analytics fall into the elementary data evaluation category. There is a tremendous value and occasions for some huge wins.
Things start getting interesting as one gets into the advanced evaluation and apply college-level math. Here one looks at correlations and regression analysis. The application of statistical techniques is to understand, not just what happened, but also why it happened. Advanced evaluation generates theories about causation facilitating to focus on what works best and scrap unproductive learning.
Predictive and Prescriptive Analytics
The most academic levels of analytics are predictive and prescriptive analytics that need graduate-level math and often rely on artificial intelligence or machine learning driven by big data sets. Predictive analytics infer, ‘based on what has happened in the past, here is what is most likely to happen next.’ Prescriptive analytics take the instance a step further and imply, ‘based on what is most likely to happen next, here is the action one should take to optimize the outcome.’ Eventually, when one gets here, they rely on highly intelligent recommendation engines that distribute just the right learning, at the right moment, in the right way to improve performance meaningfully.