Monday, 10 January 2011

Learning Analytics and Knowledge course

Jumped right into this course with a look at Hunch.

When I arrived I realised I'd visited a couple of times before - it's a link I follow from blogs and Twitter on occasion. Answer a series of picture questions and Hunch learns enough about you to make some recommendations. Keep engaging with the questions and recommendations and you should, in theory, have some useful recommendations that could help you choose what to read, where to go, what to buy. In theory. As I say, I've been there before on occasion and I have neither remembered the name nor bookmarked it. But why not?

First of all, I didn't like the options. Multiple-choice always winds me up. I hate questioners who won't give me the freedom to choose my own words and interpretations. Often all their answers are wrong, either unequivocally wrong, or wrong from my perspective.

Q. Do you live in the suburbs, a major city or a rural area? I live in a good-sized town. Skip the question.
Q. Do you tend to support liberal or conservative politicians? I live in the UK. Ask me a sensible question. Skip.
Q. Which of these sorts of fries do you prefer? Fish and chips. Oh, but the options are American fast food. Skip.
Q. Is Barack Obama a Muslim? (!) Should the site be posing as a complete idiot? Skip and End.

So the style of it all wound me up. Assuming I live in the USA wound me up. And the summary of people's responses - which is supplied each time you answer - made me feel an outsider. Only thirty-nine percent of users had identified as female (or as aliens wearing dresses, it wasn't clear), 15% had identified as Europeans. Only 17% read more than three books a month. (And did 3% really say they had PhDs? Either Hunch attracts a crowd of intellectuals, or a lot of liars.) I didn't  feel at home.

And the recommendations. Well, it spotted that I like Apple products - but after I'd answered the 'Mac or PC' question that wasn't too tricky, was it? And I said I was European, but it  recommended me five New York city museums, and five US credit cards. I was recommended a dreary series of magazines, a series of books that you'd read only if really bored, and some blogs where I didn't get beyond the first post.

Did I bookmark it? No.
Is this what I would want from a set of learning analytics? Absolutely not.

1 comment:

  1. Hi Rebecca, I've been enjoying the Hunch discussion in the moodle forum...and, of course, your comments here. The generally negative responses are quite interesting - I thought we'd get some...but not as many as we've received. Some great critical thinking happening!

    Any thoughts on what you'd look for in a learning resources recommender system? (beyond accurate results, of course). How should a system "get to know you"?