Results

In a 2x2 Analysis of Variance (ANOVA) there are two independent variables, with two treatments each. Each pair of treatments undergoes a 1-way ANOVA with the corresponding treatments of the other independent variable. In this experiment the following treatments undergo a 1-way ANOVA; Treatments 1&2, Treatments 1&3, Treatments 3&4, and Treatments 2&4.

If the null hypothesis can be rejected for both 1&2 and 3&4, then the progress bar has an interaction effect on the number of mouse clicks during this experiment. Likewise, if the null hypothesis can be rejected for both 1&3 and 2&4, then the text feedback has an interaction effect on the number of mouse clicks during this experiment.

While F is greater than F-crit for the ANOVA of treatments 1&2 and 1&3, both 2&4 and 3&4 have F values much less than F-crit (see Appendix A). Thus, the null hypothesis cannot be rejected for either independent variable and there is no statistically significant interaction effect between text and progress bar feedback methods.

Performing a 1-way ANOVA on all four treatments yields an F value = 19.2, which is much greater than the F-crit value of 2.7. Because the null hypothesis can be rejected, the T Method (Tukey's Procedure) is used to isolate the statistically different treatments. As seen in Appendix B the T Method shows that treatment 1 is significantly different than treatments 2,3, & 4 (which are not statistically different).

The statistically significant difference between Treatment 1 and the remaining treatments can also be seen in the following graph.

Mean and Standard Deviation values
	MEAN	Standard Deviation
Treatment 1	29.4	18.8
Treatment 2	12.5	1.9
Treatment 3	12.2	1.6
Treatment 4	12.4	2.1

Department of Computer Sciences Direct questions and comments to the student editorial team