Database Reference
In-Depth Information
=CHIDIST(Q,K−1)or
=CHIDIST(0.333,1).
We can implement this command in any cell of Excel; see
Figure 5.7
and the
arrow in
Figure 5.7
. (As we noted, the distribution of Q is approximated by a chi-
square distribution. And we arbitrarily picked cell C15 in which to put the command.)
Press “Enter” to obtain the
p
-value of 0.564. Therefore, we accept H0, since
0.564 > 0.05. As we have noted, this indicates that we do not have suficient evidence
(it's not even close!) to reject H0.
5.3.2
SPSS
We now present Cochran's Q test in SPSS. First, we enter the data, as shown in
Figure 5.8
.
SIDEBAR: FINALLY! SOME SPSS AND EXCEL SIMILARITIES
Fortunately, the Cochran Q test requires the data to be entered in SPSS the same way it is entered in
Excel. In a sense we're avoiding the usability problem with data entry that we had when we performed
the chi-square test in Chapter 4 (which, you may recall, needed the data to be entered as a 30 × 2 data set.)
From the “Analyze” menu, highlight the menu item “Nonparametric tests,” then
the submenu “Legacy Dialogs,” and inally the sub-submenu “K Related Samples;”
see
Figure 5.8
and the three arrows in
Figure 5.9
. Note that we have only two sam-
ples, but the Cochran Q test can handle any number, so it is listed under the more
general case of “K Related Samples.”
FIGURE 5.8
SPSS input for Cochran Q test.
Search WWH ::
Custom Search