From the results of the experiment our proposed hypothesis is neither proved nor disproved. Nonetheless, it is important to note that the results of our experiment do however tend to support rather than refute our hypothesis.
Upon casual analyzation of the results of the experiment in reference to the average link click times for each of the frame placements, it would appear that the placement of the frame at the top of a web page is the most efficient location. However, the standard deviation of link click times collected in the experiment are quite high, thus weakening the assertion that the average link click times, for each of the frame placements, will either prove or disprove our hypothesis. An ANOVA(Analysis of Variance) Single-Factor test was used to reveal whether or not there are differences on the true averages associated with the different treatments(Left, Right, Top, Bottom) of the independent variable(frame placement). Unfortunately, the results of the ANOVA test further weakened the assertion that the average link click times, for each of the frame placements, would either prove or disprove our hypothesis. From the analyzation of the standard deviation of the link click times, and the results of the ANOVA test the use of the average link click times as an argument to support our hypothesis has been severely weakened. However, it is still important to take note that the average link click times for the top frame placement is less than the average link click times of the other frame placements.
Due to the fact that the analyzation of the average link click times are inconclusive, we decided for the purpose of the experiment to calculate the median of the link click times. The calculation of the median link click times for each of the frame placements supported our hypothesis, in which the top frame placement had the lowest median link click time. As opposed to the average link click times, the median link click time is often a more accurate indicator of the results of an experiment. For example, the raw data collected includes link click times that are significantly less than or greater than the rest of the results, which in some cases have drastically decreased or increased the computed average. Therefore, by taking the median of a set of data, the outlying data that significantly deviates from the rest of the data will not effect the median, thus giving a more accurate summary of the data.
After viewing the distribution of point click times for each of the frame placements, we noticed two of the point click times deviate significantly from the rest of the point click times, for this reason we removed the two subjects that yielded these results. We then reran the ANOVA test on the remaining 22 subject’s results. Unfortunately, once again there was no statistical difference found between the 4 treatments.
In Conclusion, do to the fact that the standard deviation of the link click times are so high it is difficult to make any conclusions based on the data collected for the experiment. However, the median of the test results, leads us to believe our hypothesis will be proven under more controlled experimental conditions (See Anomalies 4.3)
4.2 Subjective Comments
Each subject filled out a survey. One of the survey’s primary purposes was to determine which of the 4 frame placements each of the subjects preferred.
The results were as follows:
4.3 Anomalies
If the randomly generated number from the web site, we used for the experiment, was removed and replaced with preset numbers that each subject would view and then select the appropriate index, then we believe the standard deviation would have been lower. Therefore, we believe that a conclusion could have been made from our results.
Continue
Return to Menu Frame Placement for Optimal Web Navigation