A simple click here will initiate a download of the "ZmedOriginal.xlsx" Excel workbook ready to work with Lertap5.

 

A click here will (should) lead to a download of a file called "Zmed-IScores.csv". This file is a "comma-separated values" file containing a header row (with item titles) followed by rows of item scores, one for each person, with each item scored (0,1) -- zero if the person/student got the item wrong, and one if s/he got the item right. (If an item response was missing, it will be coded as a zero.) csv files are often used with an R package, and/or with a data analysis system such as JASP.

 

Once the ZmedOriginal.xlsx workbook has been copied to your computer, get Lertap 5 going by opening the Lertap5.xlsm file. Once it's open, return to the ZmedOriginal.xlsx workbook and use the "Interpret" and "Elmillon" options to get results. Note the excellent reliability (coefficient alpha) and quite nice distribution of test scores.

 

As to suggested exercises, well, this was an admissions test, and as such we'd normally expect a cut-off score to be applied. The Mastery= assignment is used to set cut-off scores. What you could do is experiment with various Mastery= settings to see how they affect the pass rate (number of students passing the test), and the "Estimated number of incorrect classifications".

 

Suppose the medical school had places for just 500 students; where should the cut-off score be placed so that only 500 of the 2,470 applicants pass?

 

Try reducing the number of items and see how that impacts test reliability (coefficient alpha) and the "Estimated number of incorrect classifications". Could we bring the test down to a 70-item exam and still have adequate reliability while not raising the "Estimated number of incorrect classifications" by too much?

 

How to reduce the number of items? There are several ways, the easiest being to use a *exc line in the CCs sheet, after the *key line. You could do this rapidly by copying the *col line, pasting the copy after *key, changing "*col" to *exc" and then simply deleting some of the column numbers found in the parentheses after what will be *exc, perhaps by starting from the right end (keep the first items, delete only some of the last ones).

 

Intrigued by these questions but not sure how to go about answering them? Try dropping an email to lertap5@gmail.com.