Module 2.9 Model Misspecification and Qualitative Dependent Variables

Question 1

three broad categories of model misspecification (ways in which model can be specified incorrectly)

Answer 1

1. functional form can be misspecified
2. explanatory variables are correlated with error term in time series model
3. other time series misspecifications that result in nonstationarity

Question 2

3 types of functional form misspecification

Answer 2

1. important variables omitted
2. variables should be transformed
3. data improperly pooled

Question 3

3 types of misspecifications related to explanatory variables and correlation with error term

Answer 3

1. lagged dependent variable is used as independent variable
2. function of dependent variable is used as independent variable
3. independent variable measured with error

Question 4

dummy variable that takes on value of either one or zero

Answer 4

qualitative dependent variable

Question 5

model that is used when ordinary regression model is not appropriate for situations that require qualitative dependent variable

Answer 5

logit model

Question 6

assumes that residuals have logistic distribution which is similar to normal distribution but with fatter tails

Answer 6

logit model

Question 7

used to estimate probability of discrete outcome such as probability of default

Answer 7

logit model

Question 8

used to estimate coefficients for logit models

Answer 8

maximum likelihood method

Question 9

coefficients for logit models

Answer 9

relate independent variables to likelihood of even occurring

Question 10

interpreted as change in “log odds” of event occurring per one unit change in independent variable, all else constant

Answer 10

slope coefficient in logit model