Comparing two designs (or anything else!) using independent sample T-tests - Improving the User Experience through Practical Data Analytics - page 63

indicates they are not assumed equal. (Of course, as we saw in the Excel output, and
see here again, it does not really matter which one we use—there is no material dif-
ference between the two p -values.)
What the Levene test says is that we should use the “Sig.” (i.e., p -value) of 0.023
(the 0.0232 of Excel rounded to three digits), instead of the 0.025 (the 0.0246 of
Excel rounded to three digits). See arrow in Figure 2.11 . (As noted earlier in the
Excel section, this is not a material difference.)
FIGURE 2.11
SPSS output.