CCSE Reading group: SIGCSE Journal Club - Why Minimal Guidance During Instruction Does Not Work
Published: 2 November 2020
We will again be running our own chapter of the SIGCSE Journal Club at 13:00 before the main UK session. The reading is a paper by Kirschner et al. on minimal guidance.
When I first read this paper, it gave me a strong jolt as I realised that our principal pedagogical activity in an introductory programming class is based on minimal-guidance - we introduce very little, use few worked examples if any, and then say “go and solve these problems”. Hence, I find the arguments quite compelling.
However, when I ask other computer scientists to read this paper, I have often found they disagree strongly with it. Who is right, I wonder? One aspect I find is that my fellow readers often assume the authors’ views apply to **all** instruction. However, somewhat hidden, which is a shame, on the second column of P80, the authors refer to the expertise-reversal effect - and the indication that direct instruction is best for novices but that as one moves towards intermediate/expert, then minimal guidance becomes the best learning strategy - and this is explained with further reference to learning and the cognitive architecture which is the backbone of the paper.
We do know that many who self-select to study computer science fail to become good programmers. Many drop out completely at an early stage; and many others remain in the programme, but try to avoid programming. Somehow, they just don’t get it. What has failed them?
Looking forward to the discussion on Monday…
First published: 2 November 2020