Analyzing quantitative usability data
Typically, analyzing data from usability testing begins with using “descriptive statistics.”
The most common descriptive statistic for usability testing is frequency
counting.
Most quantitative data collection methods for usability testing (e.g.,
post surveys, recording procedures, think alouds, debriefing, etc) allow
you to count the frequency of user behavior, such as the number (or percentage)
of errors that occur on tasks and the number of users who successfully
perform tasks.
Example
Post-survey question – “Overall, do you believe this technology improved your knowledge of assessing instruction?”
| Response | Frequency | Percentage | |
|---|---|---|---|
Strongly disagree |
0 |
0% |
0% Disagree |
Disagree |
0 |
0% |
|
Neutral |
3 |
30% |
30% Neutral |
Agree |
6 |
60% |
70% Agree |
Strongly agree |
1 |
10% |
|
Total |
10 |
100% |
|
Typical user behavior is another descriptive statistic commonly used. Generally, this is used for determining the average (mean) of a behavior (e.g. time taken) for a certain task.
Example
From test monitor (data logging sheets) –
Task 4c: “Participants will find web page on how to write survey
questions for assess teaching”
| Participant | Time completed (min:secs) |
Errors |
|---|---|---|
Participant 1 |
1:30 | 0 |
Participant 2 |
4:20 |
5 |
Participant 3 |
5:20 |
3 |
Participant 4 |
4:10 |
3 |
Participant 5 |
3:32 |
4 |
Mean |
3:46 |
3 |
For greater usefulness, both of these examples should include text explanations to help readers interpret the quantitative results.
