Database Reference
In-Depth Information
FIGURE 9.8
Variables brought over to Variables dialog box; SPSS with illustrative data.
In addition, we might as well see if the r, that we ind, is statistically sig-
niicant, and if so, have it lagged. (This is essentially testing our old friend the
null hypothesis. H0 is that the true value r [often referred to as “ρ”] is 0, vs.
ρ ≠ 0. Refer to Chapter 1 for a brushup on hypothesis testing.) Again, these are the
default options in the bottom portion of the “Bivariate Correlations” dialog box
in Figure 9.7 .
Figure 9.8 shows the variables brought over to the “Variables” dialog box.
We now click on OK, and get the output shown in Figure 9.9 .
Of course, we get the same answer we got when doing the problem in Excel.
However, the SPSS output provides even more value: it tells us that the 0.895 value
of r is statistically signiicant (based on a signiicance level of 0.05), because (1) the
asterisk on the 0.895 value tells us that, and, (2) the p -value = 0.04, which, indeed, is
less than 0.05. (Recall that the p -value is always notated as “Sig.” in SPSS.)
9.3.3 CORRELATION APPLICATION TO BEHEMOTH.COM
You'll recall that Hans Blitz told you to prove to him that Behemoth spent $80 million
getting rid a feature that clients actually wanted.
 
Search WWH ::




Custom Search