[ACT-R-users] masters thesis, with code on the web on testing telephones

[Frank E. Ritter]

Similar to Wayne's email, here's a master's thesis that has just come 
out of the lab:


Freed, A. R. (2003). The effects of interface design on telephone 
dialing performance (Tech. Report No. 2003-1). Applied Cognitive 
Science Lab, School of Information Sciences and Technology, Penn 
State.

http://acs.ist.psu.edu/acs-lab/reports/freed03.pdf.

Abstract:

We look at several prototype telephone interfaces with a variety of 
tools for automatic testing. Several analyses are presented. The 
first analysis uses Fitts' Law to determine the minimum possible 
usage time. Fitts' Law provided a quick analysis of the telephones 
but did not consider any factors other than target sizes and 
distances. The second analysis uses a cognitive model to make 
predictions about other factors such as visual search. The model used 
an organized search strategy that produced similar behavior to the 
experiment subjects but failed for a few extreme interfaces. Although 
the model dialing time predictions are too high, we propose that this 
model is worth paying attention to because the dialing time orderings 
among phones and number of fixations closely follows the observed 
data.

The third analysis uses a human experiment to test the predictions 
with reaction times and eye-tracking data. The experiment subjects 
were able to easily cope with these differences and retain an 
organized strategy, but the model
was forced to revert to a random search strategy. The users often 
made errors while using interfaces but the cognitive model did not 
make the same types of errors.  Our approach should be generalizable 
to other types of interfaces, including VCRs, software applications, 
and
webpages. Our analysis techniques can eventually remove the need for 
expensive and time-consuming user testing, which may be replaced with 
analytical and cognitive model analysis.