# 13.) Dummy Dependent Variable Techniques Flashcards

A linear probability model is…

a linear-in-the-coefficients equations used to explain a dummy dependent variable

The term linear probability model comes from..

the fact that the right side of the equation is linear while the expected value of the left side measures the probability that Di = zero

What are the two problems with a linear probability model?

- ) r squared bar is not an accurate measure of overall fit.
- ) d estimated ith observation is not bounded by 0 and 1

For most researchers, the major difficulty with the linear probability model is…

the unboundedness of d estimated ith observation

The binomial logit is..

an estimation technique for equations with dummy dependent variables that avoids the unboundedness problem of the linear probability model by using a variant of the cumulative logistic function

In a binomial logit model are the d estimated ith observation now limited by 0 and 1?

Yes

The maximum likelihood (ML) is used in logits because?

It is an iterative estimation technique that is especially useful for equations that are nonlinear in the coefficients.

ML estimation is inherently different from least squares in that…

it choose coefficient estimates that maximize the likelihood of the sample data set being observed.

ML has a number of desirable large sample properties

- ) ML is consistent and asymptomatically efficient (unbiased and minimum variance for large samples)
- ) With very large samples, ML has the added advantage of producing normally distributed coefficients, allowing the use of typical, allowing the use of typical hypothesis testing techniques.

a. ) So sample sizes for logits should be substantially larger than for liner regressions, some researchers seek samples of 500 or more.

What are some of the key differences btw an equation estimated using linear probability and logit?

- ) The dependent variable in a logit equation isn’t the same as the dependent variable in a linear probability model
- ) the slope of the graph of the logit changes as Di moves from 0 to 1. Thus the change in the probability ath Di = 1 caused by a one-unit increase in an independent variable (holding the other independent variables constant) will vary as we move from Di = 0 to Di =1.

How can we interpret estimated logit coefficients? What are the three reasonable ways to interpret them?

- ) Change an average observation.
- ) Use a partial derivative.
- ) Use a rough estimate

Which approach do the authors suggest to interpret estimated logit coefficients?

They suggest that to get a rough approximation of the economic meaning of a logit coefficient, multiply by 0.25 (or, equivalently, divide by 4).

How do the estimations of linear probability and logit estimation differ?

They differ mainly in that logit does not produce Dis outside the range of 0 and 1.

The logit coefficients need to be…

divided by 4 to get meaningful estimates of the impact of the independent variables on the probability of passing the test.

The binomial probit model is…

an estimation technique for equations with dummy dependent variables that avoids the unboundedness problem of the linear probability model by using a variant of the cumulative normal distribution.