WORKING MEMORY SOFTWARE

code written by John Towse in collaboration with Graham Hitch, Zoe Hamilton, Una Hutton

This page is used to provide information about Macintosh-based computer programs written to test children's working memory and available to download. They were developed as part of a funded research grant (ESRC grant R000222789).

 

The Macintosh programs were originally written in TrueBasic and have been compiled to run as standalone applications. They should run on any PowerPC Mac set to a screen resolution of 640x480. They are known to work successfully under MacOS system 7 and system 8. They have not been tested on non PowerPC machines, and performance on these machines is not known. For details on these programs, please consult the author (John Towse, j.towse@rhbnc.ac.uk), and for experimental data from these programs, see more information in the technical reports pages (CDRG 8) with details on response norms for the sentences involved. Psychological data using the software is being prepared; please contact John Towse for more details.

 

"But isn't there working memory span software already on your other pages?" One feature constrasting the span software below with previous programs (see older pages here) is that the span tests provide a more generic test of working memory span. Previous software had been written so as to compare working memory span when the last item in a trial was a long or short processing item. The programs below sample from a stimulus pool (slightly different from before) that is the same for all sentences in an experimental session. Another difference is the criteria set for termination of test assessment, as described below.

 

 

Reading Span - Subjects complete high-Cloze sentences and attempt to remember the words produced. The program estimates the maximum number of items that can be remembered.

The program begins by offering some processing practice trials, and then proceeds to test working memory span, beginning with 2-item sets. There are three trials with these 2-item sets, and three trials at all additional set lengths up to the maximum test of 7-item sets (thus providing 5 additional trial levels). Further information below.

Reading Period - Subjects complete high-Cloze sentences and attempt to remember the words produced. The program estimates the maximum length of sentences that permit accurate recall.

The program begins by offering some processing practice trials, and then proceeds to test working memory, with 4-item sets. Provided the subjects recalls at least one of the 3-item sets correctly at the briefest level, the test continues by selecting from a new (longer) set of sentences, up to maximum of 5 additional levels. If the subject fails to recall the words correctly from each of the first three trials, the program will continue by presenting 2-item sets at increasing lengths to provide a test at a reduced level. See additional information below.

Operation span - Subjects solve arithmetic equations and attempt to remember their solutions. The program estimates the maximum number of items that can be remembered.

Analogous to the tests above, the program commences with the opportunity for practice trials with representative arithmetic equations. Then the working memory span trials begin with 2-item sets. There are three trials with these 2-item sets, and three trials at all additional set lengths up to the maximum test of 7-item sets (thus providing 5 additional span levels). Further information below.

Operation period - Subjects solve arithmetic equations and attempt to remember their solutions.The program estimates the limits in the subject's ability to remember numbers in face of progressively longer arithmetic computations.

The program begins by offering some processing practice trials, and then proceeds to test working memory with 4-item sets. Provided the subjects recalls at least one of the 4-item sets correctly at the briefest level, the test continues by selecting from a new (longer) set of materials, up to maximum of 5 additional levels. If the subject fails to recall the words correctly from each of the first three trials, the program will continue by presenting 3-item sets at increasing lengths to provide a test at a reduced level. See additional information below.

Additional information for both programs

 

The program will continue to test the subject at the next level provided at least one the three trials was recalled completely correctly (it is of course possible to subsequently score performance on the basis of stricter criteria by going over the individual output files). If two or three of the trials are recorded correctly, then a picture is displayed to congratulate the subject on their performance. The currently active mac sound alert is also called at this point (experimenters may want to choose a suitable sound file to be used here, with appropriate voice or sound).

For reading span and reading period, pressing carriage return informs the computer that the expected completion word was supplied for the sentence. If the individual's response is not the target word, then type the word itself (the computer will continue to present experimental events to the subject while the experimenter completes the supplied word). For reading span and period, typing the first three letters of the recalled word suffices to trigger a correct response.

 

To abort the programs, press the "+" character during presentation of a processing element (e.g. a card displaying a sentence). Performance is recorded in the data file after each test level is complete. The recording indicates for each trial, the material presented, the response this generated, the memory for this item produced in recall , the time to produce this word, and the estimated time to produce the recall response (though note this incorporates typing time too, and may be inaccuarte without additional verification).

 

Disclaimer: In no event will John Towse, or employers be liable to you for any consequential, incidental or indirect damages arising out of the use or inability to use the software.