Wednesday, September 12, 2007

negative > positive > residuals

this is some of what I have been doing today. thought you would be interested!? I have been looking at the residuals from the exam result from the students I taught last academic year.the image below is not the actual data as that is confidential but a simple illustration.'

The term "residual" defines the fraction of variability which cannot be explained by a given model. Residuals should be inspected carefully when establishing models, since they give valuable hints on the fulfillment of model assumptions.'
Residuals are calculated by subtracting the model estimates from the actual data. The figure below shows the model and the data on the left, and the residuals on the right.
it is interesting honest!
I found out by doing a girl / boy split, that the girls did less well gaining high grades but were more consistent against expectations. whereas the boys consistently achieved above their target. on the whole results were good, but does this now mean I have to re-consider how I teach girls?

No comments: