by Stephen Downes
Oct 20, 2014
On the Question of Validity in Learning Analytics
So your learning analytics have produced a result. How do you know you should rely on it? As Adam Cooper writes in this post, there are two dimensions of assessment of analytics results: reliability (or, how closely focused the results are on a single value), and validity (or, how closely the results are to the correct result). Note, he writes, that mere predictive accuracy is not enough to establish validity. How does the prediction compare to a random result? How many false positives and false negatives were there? The prediction could be accurate, in other words, but lucky. But more, we need to ask whether the tool could ever be used in practise and whether the results generalize or are reproducible.
This newsletter is sent only at the request of subscribers. If you would like to unsubscribe,
Know a friend who might enjoy this newsletter? Feel free to forward OLDaily to your colleagues. If you received this issue from a friend and would like a free subscription of your own,
you can join our mailing list. Click here to subscribe.