Hi,
I have a similar issue. I have been trying to replicate the WebPA calculations for peer-only assessment, particularly when students have not participated.
In a real peer assessment using WebPA Version: 2.0.0.10 [lti], for one group of four students, only two students responded:
Received
Awarded A B C D total awarded
A n/a - - - 0
B 2 n/a 3 3 8
C - - n/a - 0
D 0 2 2 n/a 4
Number awarding 2 1 2 1
Group size 4
The scores as a proportion of the total awarded by each student (who completed the peer assessment exercise) are:
Received
Awarded A B C D
A – – – –
B 0.25 – 0.38 0.38
C – – – –
D 0.00 0.50 0.50 –
Unadjusted score 0.25 0.50 0.88 0.38
Because there is no self-scoring, the students who did not complete the exercise (A and C) received two scores, while the others (B and D) received only one. Since each student should have received three scores, presumably the correction factor for students in receipt of two scores should be (N – 1)/n = 3/2 = 1.5, and for students in receipt of only one score it should be 3/1 = 3.
Using these correction factors, the WebPA scores corrected for non-responders are:
A B C D
0.38 1.50 1.31 1.13
However, the system has generated the following corrected scores (WebPA score) for these students:
A B C D
0.58 1.17 1.21 1.04
Can anyone advise how these values have been calculated?
Thank you,
Philip Smith
____________
Dr I. Philip Smith
School of Biological Sciences, University of Aberdeen
|