CSfilter: One-question test to predict if you can learn programming.

December 1, 2011 10:09 PM

A while back, there as a controversial (flamebait) blog post by a CS professor about what predicts success in programming. He had a "one-question" test that he claimed was strongly predictive. 1) The actual question was a "whats the value of variable x" after a few steps of manipulation sort of thing. 2) He interpreted it as meaning "people who get this right actually accept that the computer does exactly what you tell it to, not what you wish it to do" Help me find this post.
posted by gregglind to Computers & Internet (3 answers total)

Best answer: I think you're looking for Mental models, Consistency, and Programming Aptitude by Bornat, Dehnadi, and Simon.
posted by d. z. wang at 10:35 PM on December 1, 2011

Note, at least if I recall correctly, that success was predicted not based on who got the right answers. Rather those who answered multiple questions consistently was considered a positive result, even if they were wrong, as long as they were wrong in the same way.
posted by vasi at 11:45 PM on December 1, 2011

Best answer: Note that this tests for logical consistency, not correctness. One of the answers was indeed correct in Java, but that's not what they were testing for: the point of the test was to see whether or not the students used the same mental model to answer each question.

Also, note that it didn't really work; consistency did not predict whether students would be successful in programming ("We can see from the experiments reported above that our test doesn't work if the intake is already experienced, and in experiment 3 didn't work at all. We cannot claim to be separating the programming goats from the nonprogramming sheep: experiment 3 demolishes the notion that consistent subjects will for the most part learn well, and others for the most part won't.")
posted by vorfeed at 11:51 PM on December 1, 2011