402 Statistical Methods
Comparing Means
The ANOVA table has led you to reject the hypothesis that the mean single-
room price is the same in all four cities and to accept the alternative that
the four means are not all the same. Looking at the mean values, you might
be tempted to conclude that the high price for New York City hotel rooms
is the cause and leave it at that. This assumption would be unwarranted
because you haven’t tested for this specifi c hypothesis. Are there signifi cant
differences between the other cities as well? To fi nd out, you need to calcu-
late the differences in mean value between all pairs of cities and then test
the differences to discover their statistical signifi cance.
Excel does not provide a function to test pairwise mean differences, but
one has been provided for you with StatPlus.To create a matrix of paired differences:1 Click Multivariate Analysis from the StatPlus menu and then click
Means Matrix.
2 Click the Data Values button, and select Price from the list of range
names. Click OK.
3 Click the Categories button, and select City from the range names
list. Click OK.
4 Click the Use Bonferroni Correction checkbox.
5 Click the Output button, and direct the output to a new worksheet
named Means Matrix. Click OK.
Figure 10-8 shows the completed dialog box.Figure 10-8
The Create
Matrix
of Mean
Differences
dialog box