Will anybody buy? Logistic regression - Improving the User Experience through Practical Data Analytics - page 300

if the person views In-Depth Analysis of National Stories as very impactful in the
decision. So, how do we interpret this? Carefully, in a word. But it's safe to say that
increasing the amount of in-depth analysis of national news will not increase likeli-
hood of adoption. If folks were asked what the regular news networks such as CNN
and MSNBC News do well, many would respond that they cover national news very
well. Therefore, the negative coeficient of “national” may be relecting that this
coverage is not lacking, and is readily available for free, so if it is a person's primary
interest, he/she would not be as inclined to pay for an online news experience.
11.5.2 DUE DILIGENCE COMPARING STEPWISE RESULTS TO
REVISED BINARY REGRESSION RESULTS
It can be useful to compare the last step of the stepwise analysis to yet another (binary
logistic) regression analysis in which we enter only the ive variables that were
deemed signiicant in the original binary logistic multiple regression. This result is in
Figure 11.19 , where we've entered only the signiicant variables we identiied in the
original regression model.
FIGURE 11.19
Multiple regression analysis with only the ive signiicant variables in the multiple regres-
sion shown in Figure 11.16 ; SPSS with CharlestonGlobe.com.