Changes and enhancements made in recent months:


Added two "practical exhibits"


These "exhibits" showcase examples of Lertap5 in action, one for a cognitive test and another for an affective scale on depression. They are primarily introductory in nature but do delve a bit into more advanced topics. Both provide the chance to download copies of actual data.


The cognitive test is called "Test13" and is available here. The depression inventory (or scale) is a very popular, true-to-life scale and may be accessed via this link.


Missing data now highlighted in two ways


The first way: cells with missing data will have a background highlight in the Data sheet. This makes it very easy to get an idea of the extent and location of missing data.


The second way results when users ask Lertap5 to create a "csv" file for subsequent input to R and RMD scripts, and/or to JASP and SPSS. When Lertap5's Omega1 macro creates this file, "NA" values will be written to indicate missing data. (NA is interpreted as "value not available" in a number of scripts and apps. It's a universal missing data code.)


Read more about it here.


Updated the TAM Rasch analysis Rmd module

The TAM Test Analysis Module is a CRAN Project package described here. Rmd code modules are scripts written to work with the RStudio system.


An example of the output created by the updated Rasch-Analysis-TAM.Rmd module may be admired here, while a general presentation regarding the use of the module is here.

Documented the performance of Lertap5's Rasch item analysis

Of all the data analysis capabilities in Lertap5, the Rasch item analysis routine is unique in that it involves an iterative procedure, making substantial demands on a computer's central processing unit.  Examples of sample running times may be seen here.

Updated the "Alpha Omega Spot the Difference paper"

See it here. The paper is no longer sitting on the fence. Limitations previously found in the the JASP app have been addressed and corrected, making it the presently-recommended method for calculating coefficient omega.
The revised paper also addresses use of the "closed-form" method for calculating omega, a method now built into SPSS and Lertap5 itself, finding limitations that will rule out its use in some cases.

Improved Lertap's Rasch support

Lertap5's Rasch analysis capability has been enhanced. Now considerably easier to use, and a bit faster, it has been added as an entry in the "Run" menu (formerly it was nested within the special macros list and was not all that easy to find).




Created Lertap Version 5.11

This new version runs on both Windows and MacOS computers. (The former versions were for Windows, and for MacOS. These versions are no longer supported.)


Having now just a single version for both platforms has been made possible by Microsoft working to bring the Mac version of Excel up to par with the Windows version (or very nearly so).


Lertap 5.11 may be downloaded from this page.


Updated the set of introductory slides

Formerly they just focused on cognitive instruments (tests, exams, quizzes). Now they also include affective ones (surveys, scales, questionnaires).


They're found here (nested in a PDF file).


Added a free 4-month license option

All Lertap5 users begin their careers by downloading a copy of the free "Mini" version, something accomplished by paying a visit to this page.


The Mini version allows for processing up to 250 data records.  A free 4-month license, available to many users, will remove the 250-record limitation.  Read all about licenses at The Store.


Added plots of item correlation values for affective items

Cognitive items have had a difficulty-by-discrimination plotter for years. Now there's a plotter for affective item correlations. Rush to see examples here.


Lertap5 now has its own omega reliability estimator

The closed-form method of estimating McDonald's omega reliability figure has been added to the Windows version. Refer to Appendix D in this working paper for a discussion and example. The "Item scores and correlation" option is used to produce it.


Added a paper related to the calculation of alpha and omega reliability estimates

The paper uses SPSS, JASP, and the psych R package to derive alpha and omega estimates for 16 datasets. As has been found in other studies, alpha and omega most often turn out to be very similar in value. The paper goes on to include omega hierarchical and exemplifies its interpretation.  Despite the similarity between alpha and omega values (in most cases), the paper suggests that omega is preferred.


JASP is recommended (with reservations). Read the paper by following this link.


Increased the number-of-records limit on the Macintosh "Mini" version

The "Mini version" will now process up to 250 data records (previous limit was 100). There remains no limit on the number of items.


Users with larger data files will access the e-store to obtain a code that will entirely remove the 250 limitation.


Increased the number-of-records limit on the Windows "Mini" version

The "Mini version" will now process up to 250 data records (previous limit was 100). There remains no limit on the number of items.


Users with larger data files will access the e-store to obtain a code that will entirely remove the 250 limitation.


Updated documents having to do with using R and RStudio.

R and RStudio are free data analysis tools capable of advanced statistical analyses currently beyond Lertap5's ken.
There are special macros in Lertap5 charged with preparing R and RStudio scripts for IRT and instrument reliability work -- they will, for example, ease the process of getting ICCs (item characteristic curves) and computing coefficient omega.


A link to documentation on one of these special macros is here, while another link exemplifies how to use files prepared by the macro with R and RStudio. An example of the output produced by one of the RStudio scripts is available if you click here (results in a Word document downloading to your computer).


Assisted with the application of Lertap5 to compare exam results from sites using on-site proctoring to results collected under remote proctoring conditions during the COVID19 pandemic.

The research question in this study concerned exam integrity: did it appear that remote proctoring may have made it easier for examinees to cheat?


Exam results were analyzed using Lertap5's RSA, "response similarity analysis", with Wesolowsky's "similarity checking" method used, where possible, for corroboration.


Data were collected pre-COVID19, and during the March to May period of 2020 when the pandemic was in full stride. The pre-COVID results were collected at certified test centers with on-site proctoring. Many of these centers closed down during the COVID19 isolation period, giving rise to an increased reliance on the use of "remote proctoring".


The results from several thousand students, pre-COVID19 with on-site proctoring, did not show evidence of cheating. However, a near-equal number of examinees sitting the same exams under remote proctoring conditions did exhibit signs of possible cheating in a relatively small (but not insignificant) number of cases.



A link to the complete "updates summary" page is here.