This experiment uses a single independent variable "zoom period", with three treatments. These treatments are a zoom period of 0, 250 and 500 milliseconds (labeled treatments 1,2 and 3, respectively, in the raw data). The dependent variables map to the hypotheses presented in the preceding section. They are total time, total number of hyperlinks followed, and the number of matching answers on the post-experiment survey. The statistic used for each of these dependent variables is a one-way ANOVA with three treatments of the independent variable.
The task performed by the subjects is to navigate a Pad++ presentation to find answers to questions presented at the bottom of the screen. Figure 1 below shows the screen layout for this experiment.
Figure 1. Screen layout for experimental task
The large window on top contains the Pad++ presentation, and the smaller window on the bottom contains "Mpad", our instrumentation code that measures time to answer each question and the total number of hyperlinks followed. The three major user actions are to
The clock is stopped on each question when a carriage return is entered, and on the entire task when the last question is answered and the user clicks on the button to exit Mpad.
Retention is measured by asking the user to circle the names the 4 cities (out of 8 listed on the post-experiment survey form) that appeared on the list of main cities. The scoring of this question is as follows:
This gives a potential scale of values from -4 to 4, although in our trials there were no negative values observed.
