Database Reference
In-Depth Information
“Exactly,” you agree. You sense you're starting to reach consensus with someone
who's naturally disinclined to consensus. You plow forward.
“Even though the older folks are in the prime of their money-making years and
spend the most, our styles appeal to a younger crowd. But the good news is that the
26-45 crowd have plenty of money to burn too.”
She's warming to you. “Good point,” she says. “We're not Talbots. We're La La.
We're hip.”
“Right. And the Gen Z's are going after edgier styles.”
“But who cares?” she laughs. “They're broke!”
You laugh along. She smiles broadly.
“Well, it looks like we're right where we need to be with that home page design,”
she says, getting back to her texting. “Can you summarize in a PowerPoint for me? I
want to go to Mary with this.”
Knowing that Mademoiselle Founder Mary Myers doesn't like to wait for anything,
you put forward a tough deadline for yourself. “Sure, I'll get you that by end of day.”
“Cool!” she says. “One last thing.”
“Yes?” you ask.
“Make sure your executive summary says the customers who are buying the most
are the ones who think the new home page is most sophisticated.”
“Sure,” you reply.
“You did a great job,” she says as she leaves your ofice. “But that's all that Mary
will care about.” She pauses. “But not me. One of these days you gotta show me how
you do that stuff!”
6.10
SUMMARY
In this chapter, we have presented the F-test for testing for differences in means when
we have three or more “treatments”/columns (tasks or designs or, as here, age-groups,
or whatever we are comparing). We studied one factor, age-group in the illustrative
example, and thus, used a technique called one-factor ANOVA. It was assumed that
each age-group consisted of different people (here, obvious!!). Thus, we have indepen-
dent samples (as in Chapter 2). We illustrated its use in both Excel and SPSS.
SIDEBAR: TWO ASSUMPTIONS YOU NEED TO KNOW ABOUT
There are two assumptions (beside having independent samples) that are being made—one is that
the sophistication values in any column (age-group) are normally distributed, and the other is that,
while the average sophistication may be different for each age-group, the variability of the sophis-
tication values for each age-group is the same. As we mentioned before, these two assumptions are
somewhat robust, and modest violations of these assumptions do not materially affect the results.
However, we believe that going into detail how to test these assumptions and discussion of how to
proceed if the assumptions are materially violated is beyond the scope of the text—and, in truth, in
the vast majority of time, it won't matter anyway!!
Search WWH ::
Custom Search