Our in-class experiment used a simple activity so that everyone would be an expert in that problem domain. The experiment was to set a watch and develop models of GOMS analysis for this task. We briefly described the experiment and directions as follows:
Experiment Goal: Advance minutes by two.
Initial Condition: Watch as picture above with colon flashing.
How to operate Watch:
The watch has three buttons:
The Set minutes:
Then we formed groups to design a GOMS Task Analysis.
Model 1
Model 2 -
This models operations are the same as model 1, but their timings for a key press were derived from their observations with this particular apparatus.
Total Time for activity = 6.39 secs (based on estimate)
The purpose of this task analysis was to familiarize the class with basic GOMS analysis techniques with a simple operation. Using KLM timings was a big point of contention, so much so that one group had a similiar analysis, but used times that they thought would more accurately depict the actual time for small movements. The total time for that model was 6.39 secs. The predicted total time using KLM timings was somewhat higher than those averaged from our trial averages which was 6.03secs. We came to the conclusion that for simple operations extra cognitive time was unnecessary for all steps. People tend to learn simple operations very quickly. Another factor is that timings have to be tailored for the specific apparatus being used.
This exercise was adapted from and example by:
C. Mogensen, "GOMS analysis of watch", <URL:http://www-pcd.stanford.edu/mogens/377/goms-watch.html>.