Question

You are hired by a local election campaign to evaluate whether online education campaign materials motivated people to vote for your candidate. In order to evaluate this hypothesis, you design a survey and field it using a locally representative sample. The junior campaign strategist on your team runs a logistic regression (logit) model and presents the following table to the rest of the campaign team Table 2: Logistic Regression Model Results Coefficient Standard Error 1.851 016 .008 538 033 320 Variable Name Intercept Age (Years) Age2 (Years) Political Interest Television Exposure Newspaper Exposure Receipt of Online Education Campaign Materials 420 371 006 006 062 170 152 207 (a) Why do you think the junior campaign strategist ran a logistic regression model? (2 points) (b) Is this table a good way of displaying the results? Why or why not? (2 points) (c) Which of the factors in the e are statistically significant at the.05 level? (4 points)

Answer 1

1. This is a scenario where the response variable is of the form 0/1. Using a logistic regression we would be able to comprehend which of the factors affect our outcome. Hence he used logictic.

2. This table is moderately good in displaying the results. The ideal case will be to display the p values column as well. Also we should look for outliers test results and all. Secondly we need info of the number of data points as that would indicate robustness of model. As a preliminary check this table's fine.

3.Here we wish to test H0:a=0 vs H1: a not equals 0. Here a is the regression coefficient in each case.

The test statistic is (a-0)/s.e.(a). We shall compare it with the t distribution with n-2 degrees of freedom. As the number of data points is not given, we approximate with normality assumption here. The 97.5th quantile of N(0,1) is 1.96. Note this is a two sided test.

the value of the test statistic for each of the variables are.

here we take the absolute value of the test statistic and compare it with 1.96

age : .016/0.006 =2.67

age^2 : .008/.006=1.33

Pol. Interest = 8.68

Tel Exposure=.19

News Exposure=2.1

Receipt of study materials=2.03.

Note everything other than age^2 and television exposure is >1.96. So other than these two everything's significant.