Microsoft Office Tutorials and References

In Depth Information

Exhibit 5.24
Calculation of average (avg) and weighted average (Wt-avg)

A casual analysis of the table suggests that
Product 1
shows a 50% or greater

favorable rating for all categories. No other product can equal this favorable rating:

Product 1
is the top choice of all respondent categories except for 1

\

1 and it is

tied with
Product 3
for category 1

2. Although this casual analysis suggests a clear

choice, we can now do more formal analyses to arrive at the selection of a single

website design.

First, we will calculate the average of all
favorable ratings
for each category

\

(1

2) as a single composite score. This is a simple calculation and it

provides a straightforward method for TiendaMía.com to assess products. In Exhibit

5.24 the calculation of averages is found in F25:F28–0.6181 for
Product 1
, 0.4965

for
Product 2
,etc.
Product 1
has the highest average favorable rating. But there

are some questions that we might ask about the fairness of the calculated averages.

Should there be an approximately similar number of respondents in each category

for this approach to be fair? Stated differently, is it fair to count an individual cate-

gory average equally to others when the number of respondents in that category is

substantially less than other categories?

In TiendaMía.com’s study, there
are
different numbers of respondents in the vari-

ous categories. This can be signiﬁcant for the calculation of averages. The difference

in the numbers can be due to the particular sample that we selected. A random sam-

ple of this small size can lead to wide variation in the respondents selected. One way

to deal with this problem is to consciously sample customers to reﬂect the propor-

tion of category members that shop at TiendaMía.com. There are many techniques

and methods for formally
sampling
7
data that we cannot study here.

\

1, 1

\

2, 2

\

1, 2

\

7
Sampling theory is a rich science that should be carefully considered prior to initiating a study.

Search JabSto ::

Custom Search