3. Results

In this experiment we did not observe statistically significant differences in performance times, user satisfaction ratings nor rate of erros for individual subjects.

3.1 Performance Time

For each group (see Table 1), the performance times for the corresponding trials of all 30 subjects were compiled and the mean and standard deviations were computed. Table 2 shows these results. In order to determine if the independent variable, history styles, had a significant effect on performance time, a 3x1 (single factor) analysis of variance was computed. The ANOVA determined that the different types of history styles had no statistically significant difference on performance times at the alpha = 0.05 level. This means that there was no single history style that gave any better performance time over another.

Table 2

Click here to see a graph of performance times.

3.2 User Satisfaction

The subjective satisfaction ratings were also compiled and analyzed. The means and standard deviations were calculated for all 10 scores for each cell of Table 1. Table 3 shows the results. As with performance time, a 3x1 single factor analysis of variance was used for each treatment to detect if the history styles significantly affected users' satisfaction. The ANOVA showed that there was no statistically significant affect at the alpha = 0.05 level, indicating that one history style was not preferred over another.

Table 3

Click here to see a graph of user satisfactions.

3.3 Rate of Errors

We calculated the number of questions each subject in each treatement answered incorrectly. The means and standard deviations were calculated for these errors. A 3x1 single factor analysis of variance was used for each treatment to detect if the history style significantly affected the number of errors made by the subject. The ANOVA showed that there was no statistical significance affected at the alpha = 0.05 level. This means that one history style is not better at preventing errors over another. It should be noted that each question that was answered incorrectly received a penalty of 2 * Performance Time.

Table 4

Click here to see a graph of user satisfactions.

3.4 Memory Recall

We did not analyze the data to determine statistical significance in memory recall. However, take a look at the results to see the percentage of pages recalled per treatment.


Top | Credits | Introduction | Literature Review | Experiment | Results | Discussion | Conclusions | Acknowledgements | References | Appendices

Last modified: May 11, 1998