Make your own free website on Tripod.com


« November 2007 »
S M T W T F S
1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30


Some URLs
Main Home Page
My Research Journal
Wednesday, 7 November 2007
Data analysis
Mood:  chillin'
Now Playing: Incessant noise of the Jennie Lee Building
Topic: Data Analysis

Well, I've done some preliminary data analysis with the 22 participants I have, there seems to be two outliers (J&E) which seems to be influencing the data unduly so I may have to remove them in the final analysis - but we shall see what happens.

With the Latin square design, thankfully it seems as if the order does not make a difference to how the problems are answered, however, problem 2 is being done significantly better by students than problem 1 and 3. There is no clear indication that any of the boxes are actually helping the students unfortunately - although I am still hoping something will happen in the next 14 data collection which would show something ... so far the things that I'm looking out for it seems that the students are doing better in the constructive problems using the black-box than the open and glass box ... but this is only in the marks they're receiving - I still have to see what is happening in terms of self-explanations ... hopefully there is something there which might be able to help. Further, in high conceptual stuff, the glass box seems to be doing the worse, but in general everyone is doing worse with this stuff, so it might be as Doug's say a glass-bottom effect. I think the reason why students are getting the high conceptual parts wrong in the LP is because they're drawing from real life experience which is interesting as this usually helps people in getting the answer correctly in most problems but as James' says it is because they're not seeing or formulating it as a linear programming problem in their minds.

However, I think for the high conceptual it is dependent on part on some low conceptual in the interpretive questions and the procedural in the constructive questions, probably should look when they get the 1st part correct, do they go on and do the high conceptual part right ... which I think might be more interesting. 

Hopefully, when I come to the self-explanations I can see something happening which might explain the differences between the boxes (hopefully!)


Posted by prejudice at 11:48 AM GMT

View Latest Entries