Ch12 Verified Test Bank Dummy Dependent Variables

Chapter 12

True or False Questions:

1. True or False: A dichotomous dependent variable signifies that an event either happened or did not.
2. True or False: Both OLS and probit models require that the error term be normally distributed.
3. True or False: Maximum Likelihood Estimation (MLE) uses t-tests, just like OLS.
4. True or False: We can use OLS to estimate a LPM model.
5. True or False: Probit and logit coefficients are interpreted the same way as LPM coefficients.

Multiple Choice Questions:

1. Which of the following is an appropriate way to interpret a coefficient on a continuous independent variable (X1) from a probit model?
   1. Calculate the difference in fitted values when the variable is at its actual value and increased by a standard deviation, holding all other variables at their actual values.
   2. Standardize each observation by dividing each observation by the standard deviation.
   3. The coefficient indicates how much a one unit increase in X1 changes the predicted probability.
   4. Use a latent variable.
2. Which of the following is an appropriate way to interpret a coefficient on a continuous independent variable (X1) from a LPM model?
   1. Calculate the difference in fitted values when the variable is at its actual value and increased by a standard deviation, holding all other variables at their actual values.
   2. Standardize each observation by dividing each observation by the standard deviation.
   3. The coefficient indicates how much a one unit increase in X1 changes the predicted probability.
   4. Use a latent variable.
3. In order to run a hypothesis test on multiple coefficients to check if they are different from one another (equal, bigger or smaller) in a probit model, we:
   1. Use an F Test
   2. Use a likelihood ratio test
   3. Run multiple t-tests
   4. Use a Chi-squared test
4. Which of the following is the equation for a likelihood ratio test?
   1. LR=(logL_ur - logL_r)
   2. LR=2(logL_ur - logL_r)
   3. LR=(logL_r - logL_ur)
   4. LR=2(logL_r - logL_ur)
5. In a probit model, the interpretation of the estimated effect of X1 on the probability Y=1 depends on:
   1. The current level of X1
   2. The current level of the other independent variables.
   3. The current level of Y.
   4. Both a and b
6. Which of the following is a characteristic of latent variables?
   1. The value of the latent variable is high, than the dependent variable for that observation is likely to be 0.
   2. The value of the latent variable is high, than the dependent variable for that observation is likely to be 1.
   3. It is an observed continuous variable reflecting the propensity of an individual observation of Yi to be equal to 0 or 1.
   4. They are normally distributed.
7. The cumulative distribution function
   1. Tells us how much of a normal distribution is to the right of any given point.
   2. Tell us how much of a normal distribution is to the left of any given point.
   3. Shows the probability for each possible value of the random variable.
   4. Has the same shape as a normal distribution, but wider tails.
8. When is a probit model preferred to a logit model?
   1. Use probit when the independent variables are whole numbers
   2. Use probit when the dependent variable ranges from -1 to 1.
   3. Use probit in all cases since it gives slightly more accurate results than logit.
   4. There is generally no clear reason to pick one over the other as models produce similar results.
9. Which of the following is false?
   1. Fitted values from probit models and logit models are very similar when they are based on the same data.
   2. The coefficients in a probit and logit model are very similar when they are based on the same data.
   3. Probit models are slightly more accurate.
   4. Logit models make sense when using logged variables.
10. Which of the following is not a property of MLE if there is no endogeneity?
    1. Parameters are normally distributed.
    2. Parameters are consistent.
    3. Fitted values can be produced.
    4. Coefficients minimize the sum of squared residuals.
11. Which of the following is a drawback of using LPM models?
    1. LPM models cannot handle dichotomous dependent variables.
    2. LPM fitted values do not always fall within the range of 0 and 1.
    3. LPM models require us to assume errors are normally distributed.
    4. LPM models are complicated to interpret.

1. Write down the equation for a probit model with one independent variable.

1. Write down the equation for a logit model with one independent variable.

1. List and explain the benefits and drawbacks of employing a linear probability model.
2. Explain the logic behind the use of latent variables in order to explain observed variables.

1. Explain how we can interpret probit coefficients using the observed-value, discrete differences method in the case where X1 is continuous.