Microsoft Office Tutorials and References
In Depth Information
have a count of 14 females (marginal column totals
14). By not includ-
ing a dimension, such as Age, we ignore age differences in the data. The same is
true for Income. More precisely we are not ignoring Age or Income, but we are
simply not concerned with distinguishing between the various categories of these
two dimensions.
So what are the preliminary results of the cross-tabulation that is performed in
Table 5.3 and 5.4? Overall there appears to be more good than bad evaluations of
Product 1, with 11 bad and 18 good . This is an indication of the relative strength of
the product, but if we dig a bit more deeply into the data and consider the gender
preferences in each region, we can see that females are far less enthusiastic about
Product 1, with 7 bad and 7 good. Males on the other hand, seem far more enthu-
siastic, with 4 bad and 11 good . The information that is available in Table 5.4 also
permits us to see the regional differences. If we consider the South region, we see
that both males and females have a mixed view of Product 1, although the number
of respondents in the South is relatively small.
Thus far, we have only considered count for the data field of the cross-tabulation
table; that is, we have counted the respondents that fall into the various intersec-
tions of categories—e.g. two Female observations in the West have a bad opinion of
Product 1. There are many other alternatives for how we can present data, depending
on our goal for the analysis. Suppose that rather than using a count of respondents,
we decide to present the average income of respondents for each cell of our data
area. Other options for the income data could include the sum, min, max, or standard
deviation of the respondent’s income in each cell. Additionally, we can calculate the
percentage represented by a count in a cell relative to the total respondents. There
are many interesting and useful possibilities available.
Consider the cross-tabulation in Table 5.3 that was presented for respondent
counts. If we replace the respondent count with the average of their Income, the data
will change to that shown in Table 5.5. The value is $100,750 for the combination of
Male/Bad in the cross-tabulation table. This is the average 4 of the four respondents
found in Table 5.2: #8–$123,000, #10–$48,000, #16–$138,000, and #27–$94,000.
TiendaMía.com might be very concerned that these males with substantial spending
power do not have a good opinion of Product 1.
...
3+4+2+5
=
Table 5.5 Cross-tabulation
of gender and product 1
preference in terms of
average income
Product 1
Gender
Bad
Good
Totals
Female
61,571.4
25,071.4
43,321.4
Male
100,750.0
47,000.0
61,333.3
Totals
75,818.2
38,472.2
52,637.9
4 (123,000 + 48,000 + 138,000 + 94,000) / 4
=
$100,750.
 
Search JabSto ::




Custom Search