What is an R. Count in the item analysis?

What is an R. Count in the item analysis?

Patrick Malley - келді
Number of replies: 5
I know what the function of an R. Count is - to count the total number of students who guess a particular answer.  But, what is "R"? 

This is probably something simple or statistical that I am missing?  Is it ratio?
In reply to Patrick Malley

Re: What is an R. Count in the item analysis?

Joseph Rézeau - келді
Core developers қатысушының суреті Particularly helpful Moodlers қатысушының суреті Plugin developers қатысушының суреті Testers қатысушының суреті Translators қатысушының суреті
Maybe simply R = Right (i.e. total number of students who guessed the answer correctly) ?
In reply to Joseph Rézeau

Re: What is an R. Count in the item analysis?

Tim Hunt - келді
Core developers қатысушының суреті Documentation writers қатысушының суреті Particularly helpful Moodlers қатысушының суреті Peer reviewers қатысушының суреті Plugin developers қатысушының суреті
I have not worked this out yet. Looking at the quiz reports is pencilled into my personal roadmap for between October and December.

It may be some sort of regression co-efficient. Other quiz analysis systems use a statistic that is basically "is there a good correlation between the students' score on this question, and their score on the whole quiz?" If there is a good correlation, then that question is probably a good indicator of ability.

By the way, have you tried clickin on the help buttons or looking in MoodleDocs?
In reply to Tim Hunt

Re: What is an R. Count in the item analysis?

Maria B. - келді
Good indicator of reliability is "image" - squared multiple correlations between each variable and all other variables.

Discrimination Coefficient means Corrected Item-Total Correlation or not?

DC = Sum(xy)/ (N * sx * sy)

Some statistical packages, like SPSS, use overall quiz scores decremented by score for that particular queastion i.e. C = Y - X



In reply to Maria B.

Re: What is an R. Count in the item analysis?

Timothy Takemoto - келді

I don't think it is corrected. It does not mention that in the documentation.

When all the students get a question correct, and so I guess that the discrimination coefficient should be 0 (zero) because it has no power of discrimination but it is also not reverse discriminating, but in moodle 1.53 it is calculated to be 999.

Tim