The actual multiple-choice items used in this very basic 13-item mathematics test may be seen here.

 

An Excel workbook with item responses from 2,976 students, "Test13.xlsx", may be downloaded by clicking here. It includes a "zone" code for each of the students which indicates if they were from the north, south, east, or west of the country. The zone code can be used to breakout score and item responses using two standard Lertap5 options (breakout test scores, and breakdown item responses).

 

A data subset with responses from the 2,727 students who gave answers to at least 8 of the 13 items is available here. This subset excludes 249 students who failed to answer 5 or more of the 13 items. Unanswered questions are coded as 9 in Test13 data.

 

A presentation/discussion of Test13 items and an analysis of results, suggested as perhaps a useful reference for use with test and measurement classes is here.

 

A comma-separated values ("csv") file of item scores may be downloaded from this link. This file has responses from those 2,584 students with test scores in the range of 2 to 12 (students with extreme scores of 0, 1, and 13 have been omitted). This file is meant to be used with R and RMD scripts, such as Rasch-Analysis-Tam.Rmd. The name of the file, Omega-IScores.csv, relates to this topic.  The Excel workbook corresponding to this subset of 2,584 students may be found here.

 

An item analysis report created by using the above-mentioned script, Rasch-Analysis-TAM.Rmd, is found in this pdf file; the data used in the analysis came from the 2,584 students having test scores ranging from 2 to 12.

 

The original source for Test13 is "A Course on Test and Item Analysis" by Prof. Margaret Wu. Chapter 7 introduces Test13 and discusses how to access a DAT file having the original item responses and associated data. The file's name is Maths13.dat.

 

Comments on using R with Lertap5 are found here. Lertap5 has its own Rasch module; read about it here.

 

Two of the 13 test items performed somewhat poorly, having weak discrimination. This Excel workbook is set up with two Lertap5 subtests, one with all 13 items, and one with just the best-performing 11 items. Running Lertap5's Interpret and Elmillon options will create Stats1b and Stats2b worksheets which may be compared; a look at the Stats1f and Stats2f worksheets is also worthwhile (compare the reliability and standard error of measurement values seen in the two worksheets - results are a bit better without the two weak items).