and column percentages are useful in identifying the underlying association once statistical
significance is found. This fact is underlined by the XL Data Analyst, which provides row
and column percentages only if the cross-tabulation relationship is significant at the 95
percent level of confidence.
4. The null hypothesis is omnipresent in associative analysis tests, and it is the foundation for
practically all statistical tests. We recommend that Instructors continually remind students of
the null hypothesis of no association as they review the various associative analysis tests. It
may be worthwhile to remind students that the null hypothesis is present in statistical
inference tests such as t-tests (no difference between the means of the two groups), or
analysis of variance (no difference between any two group means). For Instructors’
information, the null hypothesis concept is emphasized in the next chapter, particularly with
descriptions of bivariate and multiple regression analyses.
5. For Instructors who want their student to actually compute statistics, consider end-of-chapter
questions numbers 14 and 15. Number 14 provides observed frequencies in a cross-tabulation
along with the computed Chi-square value. It has only four cells. The answer to question 14
below has the XL Data Analyst cross-tabulation output with the row and cell percentages, so
Instructors will be spared the requirement to compute them. Question 15 will require students
to set up the cross-tabulation table and to determine the observed frequencies, plus do all
6. For Instructors who want to emphasize the scatter diagram interpretation of a correlation,
consider using Application Question 16. The correlation matrix for all five variables is
provided in the answer to question 16 below as are all the scatter diagrams. Students should
be able to build the data set in Excel for Windows quickly, and they can have Excel create all
possible scatter diagrams.
7. There are many nuances to regression analysis not treated in this chapter’s introduction to the
topic. The intent is to describe the basic concepts and to have students identify their related
values on a printout. Instructors should be aware that in no way will students become more
than fundamentally knowledgeable about regression, running it and interpreting the findings.
8. A constraint in Excel is that it will allow no more than 16 independent variables when
performing regression analysis. When students (or instructors) are confronted with a larger
number of independent variables, we recommend the following approach.
a. Identify groups of independent variables such as demographics, lifestyle
characteristics, usage and/or behavioral variables, performance evaluations and/or
satisfaction variables, etc.
b. Use regression analysis to identify the statistically significant independent variables
in the first group (such as usage and/or behavioral variables).
c. Add the next group to these variables (such as lifestyle characteristics) and determine
the statistically significant independent variables. (It is necessary to rerun the