Comparing two designs (or anything else!) using independent sample T-tests - Improving the User Experience through Practical Data Analytics - page 69

SIDEBAR: CONFIDENCE INTERVALS FOR THE DIFFERENCE IN MEANS
There is actually a formula for a conidence interval for the difference of two means when you have
independent samples. However, this formula is not available in Excel or SPSS.
With this formula, you look at the resulting conidence interval and know to accept H0 when
the conidence interval contains the value 0, and to reject H0 when the conidence interval does not
contain 0.
The formula exists on pages 70-71 of Sauro and Lewis (2012) .
2.9 SUMMARY
In this chapter, we have presented the t-test for two independent samples. This test
deals with comparing two sample means when, for our purposes, the two samples
are composed of different people and, hence, are independent. We described the cir-
cumstances when the test should be used, discussed several situations where it can be
used in the world of UX research, and illustrated its use in both Excel and SPSS. We
ended the chapter with a cautionary tale and illustration concerning an incorrect use
of conidence intervals to duplicate the results of the t-test for independent samples.
2.10 EXERCISE
1. Consider the case of the time it takes to perform a task and testing whether
the mean time it takes for one task (T1) is the same as mean time it takes for
another task (T2). Data for the times for each task are in data iles in Excel (ile
name: Chapter 2.Exercise 1.data) and in SPSS (ile name: Chapter 2..Exercise
1.data). The output in Excel is on sheet 2 and sheet 3 (respectively, assuming
equal variances and not assuming equal variances). In SPSS the output is in a
ile (ile name: Chapter 2..Exercise 1.output). All iles are on a Web site indicated
in the preface to the text.
Test whether the mean times for T1 and T2 are equal or not.
A word ile (ile name: Chapter 2.Exercise 1.discussion) is also provided, which
discusses the results.
REFERENCE
Sauro, J., Lewis, J.R., 2012. Quantifying the User Experience: Practical Statistics for User
Research. Elsevier.