15.2 Exercises

  1. Hybrid Cars (***): Consider the problem of choosing whether or not to purchase a hybrid vehicle (e.g., the Toyota Prius, Honda Civic Hybrid, Ford Escape). As an analyst, you assume that whether or not an individual purchases a hybrid depends upon the current price of gasoline (\(gas\)), the difference in purchase price of a hybrid vehicle compared to a comparably equipped vehicle (\(increment\)), college education which is represented by a dummy variable that equals 1 if individual has completed college and equals 0 otherwise (\(college\)), and a dummy variable that equals 1 if the individual is a member of an environmental organization, e.g., Nature Conservancy, National Audubon Society, (\(env\)). Answer the following questions contained in hybrid:

    1. Provide summary statistics (i.e., mean, minimum, and maximum) for each variables.

    2. Estimate a regression model that allows to calculate the probability that a person would buy a hybrid.

    3. Using your parameter estimates, compute the probability that the following ``types’’ of individuals will buy a hybrid:

      • Type I: gasoline price = 2.50; difference in purchase price = 1,500; college = 0; and member of an environmental organization = 0
      • Type II: gasoline price = 3.50; difference in purchase price = 500; college = 1; and member of an environmental organization = 1
      • Type III: gasoline price = 3.00; difference in purchase price = 1,000; college = 1; and member of an environmental organization = 0
    4. Given the above probabilities, calculate the marginal effect that gasoline prices have on the probability that each of the three ``types’’ of individuals will purchase a hybrid vehicle.

  2. EV Data (***): Using the evdata, estimate a logit model which calculates the probability that a consumer would purchase a hybrid (choice=2), plug-in hybrid (choice=3), or electric vehicle (choice=4) as opposed to a gasoline car (choice=1). In a first step, you should create a new variable that is equal to 0 if the person purchases a gasoline vehicle and is equal to 1 for everything else. Include the following independent variables in your model: \(age\), \(numcars\), \(politics\), \(female\), \(edu\), and \(income\).

  3. Voting Behavior (***): We are interested in the determinants of voting based on data from the General Social Survey (GSS). The following variables are hypothesized to be a determinant of voting behavior: \(age\), \(gun ownership\) (yes=1), \(stance on death penalty\) (in favor=1), and \(education level\) (higher number is associated with a higher education level). Run a probit and logit model and report the results. Calculate the marginal probabilities.

  4. Fatal Car Accidents (***): Research suggests that black motorists are driving slower during the day than at night because police officers could engage in racial profiling more easily during the day than at night (Kalinowki et al., 2021). Assuming that lower speeds lead to fewer accidents, the aforementioned hypothesis can be tested via data from the Fatality Analysis Reporting System (FARS). Use the data in fatalcity and fatalstate to run a logit model with \(black\) as the dependent variable and \(rain\), \(daylight\), and \(year\) as the independent variables. In addition, use \(fips\) and \(state\) as independent variables as well but use the function factor() to include them, e.g., +factor(state). Based on the two separate regression models, what do you conclude?

  5. Biking (***): Consider the data in hhpub. Your dependent variable will be called \(bike\) and indicated whether respondents in the survey use a bike or never at all. For this question, you will focus on the following variables: \(bike\), \(hbppopdn\), \(hhfaminc\), \(hhvehcnt\), \(homeown\), and \(urbrur\). You must read the codebook for this question and learn how the variables are coded. Go to the codebook and pick “Household” as the dataset (drop down menu). Estimate a probit model with \(bike\) as the dependent variable and all other variables mentioned above as the independent variables. Interpret your results.

  6. Organic Fruit Purchases (***): The data in fpdata contains survey results on how many organic tomoatoes and organic strawberries a household purchases. For the