Page 173 - Intermediate Statistics for Dummies

13_045206 ch08.qxd 2/1/07 9:58 AM Page 152
152
Part II: Making Predictions by Using Regression
Movie Enjoyment (Yes or No Data) Based on Age
Table 8-1
Age
3
10
3
4
15
4
16
3
3
18
3
2
20
2
3
4
25
2
30
4
2
5
35
1
6
1
40 Enjoyed the Movie Total Number Sampled
45 0 3
50 0 2
General Steps for Logistic Regression
The basic idea of any model-fitting process is to look at all possible models
you can have under the general format and find the one that fits your data
best. The general form of the best-fitting logistic regression model is
/ e b 0 + b x /
1
p = b x , where p is the estimate of p, b 0 is the estimate of β 0 , and b 1 is
1 + e b 0 + 1
the estimate of β 1 (from the previous section). The only values you have a
choice about to form your particular model are the values of b 0 and b 1. These
values are the ones you’re trying to estimate through the logistic regression
analysis.
To find the best-fitting logistic regression model for your data, complete the
following steps:
1. Run a logistic regression analysis on the data you collected (see the
section “Running the analysis in Minitab” for these instructions.)
2. Find the coefficients of constant and x, where x is the name of your
explanatory variable.
These coefficients are b 0 and b 1, the estimates of β 0 and β 1 in the logistic
regression model.