Hello everyone. I have tried to find any postings regarding the changes in the item analysis between 1.9 and 2.3. I have not been totally successful finding what I am looking for and was hoping someone may be able to help. Our school recently moved to 2.3 and the statistics report is great but has more info then before and it appears that the Discrimination categories are valued differently. We are use to the -1 to 1 range and are noticing that the data is now using percentage ranges etc. I am looking for a link, video or site (other than the formula breakdown...that will help explain that and some other minor questions. Any help or suggestions is appreciated.
http://docs.moodle.org/25/en/Quiz_statistics_report links through to http://docs.moodle.org/dev/Quiz_statistics_calculations. That gets rather technical, but it sounds you are looking for a technical explanation.
Oh, I just remembered this other document: http://labspace.open.ac.uk/file.php/3484/Brief_Guide_to_iCMA_reports.pdf
When the statistics were first implemented, we (well, my colleage Phil Butcher) did some careful checking. So, I am pretty confident it is right, but it would not hurt for more people to check it more carefully if they have concerns. If you instructors do identify any bugs, please let us know.
Thank you
I am reasonably confident that it is correct as is. But like you, I am open to someone demonstrating that the computation is incorrect.
My questions to them (which they have not answered yet) are:
Do you think the calculation is incorrect? Or, do you think the formula is incorrect?
Your comments are appreciated.
I am still having a challenge with this. I explained to a couple instructors I need more info about what they believe is not correct or what their real concern is.
I broke it down to: do you think the formula is wrong, do you think their is a computation error in MOODLE or do you just believe that because the data appears different than you are use to, that it concerns you?
I also mentioned, if they believe there is a technical glitch, our MOODLe host can evaluate the site. If we believe that the formula is incorrect, we need to find where it is wrong and why?
My other thought is...if we as a school are not comfortable with using this analysis, then we need to find a different way to analyze the data which could include doing it outside of moodle or paying for a customization.
Anyone have any input?