Shore '00: Student HCI Online Research Experiments

University of Maryland

Abstract
Introduction
Experiment
Results
Discussion
Conclusions

Acknowledgements
References
Appendices
Credits
Feedback

Back To Main

Effects of Splitting Text into Multiple Columns

Results

Raw data is available for download in the Microsoft Excel format from the Appendices page.

Charts & Graphs


The graphs above display the results of completion times collected during the experiment.  The first chart on the left shows the mean completion times in a bar-chart format.  As can be seen from that graph, the mean completion times vary for different treatments.  However, due to large standard deviations, the results are inconclusive.  The second graph shows line-plot of mean completion times depending on the width of the window.



These graphs display mean completion times in bar-chart and line-plot formats.  The important fact to note is the magnitude of standard deviation lines on the bar-chart plot.  That fact indicates that the mean number of errors values do not show statistically significant differences.  T-test and ANOVA are used below to test that assumption.



The plots of mean subjective satisfaction are displayed above.  These plots show some interesting results.  The bar-chart shows that mean subjective satisfaction values for three-column display are higher than those for single-column.  The line-plot also confirms that fact and shows that the mean subjective ratings decline for the single-column display as the width of the window increases.

T-tests

  Time to Completion
600 pixels 800 pixels 1000 pixels
1 column 3 columns 1 column 3 columns 1 column 3 columns
Mean 197.81 202.21 229.98 295.82 292.25 261.94
Variance 14109.87 8534.11 12977.01 40923.48 12316.95 5268.14
Observations 10 10 10 10 10 10
t Stat -0.17 -1.75 0.83
P(T<=t) one-tail 0.43 0.06 0.21
t Critical one-tail 1.83 1.83 1.83
  Number of Errors
600 pixels 800 pixels 1000 pixels
1 column 3 columns 1 column 3 columns 1 column 3 columns
Mean 1.2 0.4 0.6 0.7 1 1.3
Variance 3.96 0.49 0.49 1.57 1.78 3.79
Observations 10 10 10 10 10 10
t Stat 1.08 -0.20 -0.41
P(T<=t) one-tail 0.16 0.42 0.35
t Critical one-tail 1.83 1.83 1.83
  Subjective Rating
600 pixels 800 pixels 1000 pixels
1 column 3 columns 1 column 3 columns 1 column 3 columns
Mean 5.1 6.2 5 6.9 3.6 6.7
Variance 5.21 5.51 6.67 4.54 2.27 1.57
Observations 10 10 10 10 10 10
t Stat -0.87 -1.85 -4.72
P(T<=t) one-tail 0.20 0.05 0.00
t Critical one-tail 1.83 1.83 1.83

The tables above show the results of t-test analysis performed on the data.  The analysis was performed using Data Analysis Toolkit in Microsoft Excel 97.  The t-test analysis was performed for three different window widths (600, 800, and 1000) for every dependent variable.

ANOVA

  Time to Completion - Summary
600 pixels 800 pixels 1000 pixels
1 column 3 columns 1 column 3 columns 1 column 3 columns
Count 10 10 10 10 10 10
Sum 1978.1 2022.1 2299.8 2958.2 2922.5 2619.4
Average 197.81 202.21 229.98 295.82 292.25 261.94
Variance 14109.87 8534.11 12977.01 40923.48 12316.95 5268.14
  Time to Completion - ANOVA
Source of variation SS df MS F P-value F crit
Between groups 93689.79 5 18737.96 1.19 0.32 2.39
Within groups 847166.00 54 15688.26      
Total 940855.79 59        
  Number of Errors - Summary
600 pixels 800 pixels 1000 pixels
1 column 3 columns 1 column 3 columns 1 column 3 columns
Count 10 10 10 10 10 10
Sum 12 4 6 7 10 13
Average 1.2 0.4 0.6 0.7 1 1.3
Variance 3.96 0.49 0.49 1.57 1.78 3.79
  Number of Errors - ANOVA
Source of variation SS df MS F P-value F crit
Between groups 6.33 5 1.27 0.63 0.68 2.39
Within groups 108.6 54 2.01      
Total 114.93 59        
  Subjective Rating - Summary
600 pixels 800 pixels 1000 pixels
1 column 3 columns 1 column 3 columns 1 column 3 columns
Count 10 10 10 10 10 10
Sum 51 62 50 69 36 67
Average 5.1 6.2 5 6.9 3.6 6.7
Variance 5.21 5.51 6.67 4.54 2.27 1.57
  Subjective Rating - ANOVA
Source of variation SS df MS F P-value F crit
Between groups 78.63 5 15.74 3.66 0.01 2.39
Within groups 231.9 54 4.29      
Total 310.58 59        

These tables show the results of single-factor ANOVA analysis performed on the experimental data.  Single-factor ANOVA analysis was performed separately for every dependent variable for all the treatments.  The analysis was done using Data Analysis Toolkit in Microsoft Excel 97.



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

University of Maryland