Exploring Psychology Homepage
         
 

Sperling Test Programming Notes - Andrew Linn

This was a fun project. For some reason this particular programming task struck a chord with me. I could easily grasp the concepts presented and understood the mechanisms involved, and so was able to (in my opinion) create a better program. The more I worked on the projects involved with this grant, the more I realized the importance of having the programmers intimately involved with the subject matter.

We still have all the issues of timing as discussed in Computer Based Testing: A Question of Time, then there is the added interface problem of requiring the subject to type in the correct letter for each row/column. One of the premises is that memory of what you have seen in the flash breaks down as a function of time, so requiring people to first click on, then type in the correct letter in the correct column and correct row might introduce other factors into the data, like keyboard familiarity or just computer familiarity in general.

It might also be true that the introduced factors cancel each other out. I tend to be able to pick three correct choices per test, whether whole or partial. That matches the expectations of Sperling's test, so the point might be moot.

Once again, it came down to a usage of lists. Of the entire alphabet, only 19 letters have been deemed suitable for use (A, E, I O, Q, U and Y are omitted). Of these 19, only 12 are used per test, and they have to be presented in random order with no repeats. For each test, create a list of acceptable letters, count the list, randomly select a number between 1 and the number of items in the list, remove from the list the letter that occupies that position and add it to the list of items show to the subject. When you have 12, stop selecting letters.

Deciding how to present the output was interesting. I decided to allow the researcher to give the subject a number of practice tests. I decided to not record the results of these tests - they are practice after all. I didn't record the number of practice test given, either, which might be a piece of information someone would want.

Original Director 8 source code if you have questions, please e-mail me at: alinn@willamette.edu

 

 
Site Map
Contact Us

 

Exploring Psychology Site

Background