Comparing two designs (or anything else!) using independent sample T-tests - Improving the User Experience through Practical Data Analytics - page 55

You check your latest survey results and see you have a sample size of 18. That
is, 18 people have rated Design 1 on a 1-5 Likert scale, where you ask respon-
dents to rate agreement or disagreement with statements like: “This is a sophisticated
design,” and “1” represents “strongly disagree,” and “5” represents “strongly agree.”
(The 5-point scale is: Strongly Disagree, Disagree, Neutral, Agree, Strongly Agree).
You also have collected a sample size of 20 data points on Design 2; i.e., these 20
people saw Design 2 and responded to the same statement and using the same scale.
The data is in Table 2.1 .
The (sample) means for the ratings of the two designs (columns of Table 2.1 ) are
4.22 and 3.40, respectively.
Again, let us remember that there are 38 different people in this study. We are
testing the hypotheses:

H0:μ1=μ2
H1:μ1≠μ2
where μ1 = the true mean of the sophistication level with Design 1, and μ2 = the
true mean of the sophistication level with Design 2. And we believe it is worth

Table 2.1 Illustrative Data from a 5-Point Scale

Design 1 Design 2

4 4

5 5

5 5

2 2

4 2

4 3

4 2

5 4

5 2

4 4

5 2

5 4

5 4

4 5

5 5

2 2

4 4

4 4

3
2