Download presentation
Presentation is loading. Please wait.
Published byChester Cobb Modified over 9 years ago
1
Copyright © 2014 McGraw-Hill Education. All rights reserved. No reproduction or distribution without the prior written consent of McGraw-Hill Education. Chapter 12 Limited Dependent Variables
2
12-2 Learning Objectives Estimate models with binary dependent variables Estimate models with categorical dependent variables
3
12-3
4
12-4 Binary Dependent Variables A limited dependent variable is a dependent variable whose range of possible values is restricted in some way. A binary dependent variable is a dependent variable that only takes on two possible values: 1 if the observation possesses a certain characteristics and 0 otherwise.
5
12-5 Binary Dependent Variables Example: Q:Do you have employer-provided health insurance? 1=Yes 0 = No Data: 38,659 individuals drawn from the National Health Interview Survey
6
12-6 Linear Probability Model The linear probability model is an ordinary least square (OLS) regression model with a binary dependent variable.
7
12-7 Linear Probability Model The linear probability model is an ordinary least square (OLS) regression model with a binary dependent variable. The issue: OLS estimates values for the dependent variables that are less than 0 and greater than 1 when such values cannot actually occur. Estimated but cannot actually occur
8
12-8 A More Preferred Estimator A more preferred estimator in the case of binary dependent variables would be one that did not estimate values less than 0 or greater than 1 and that allows the marginal effects to change as the values of the independent variables change.
9
12-9 The Logit Model The logit model is a popular specification for estimating population regression models with binary dependent variables. Notes: The Logit model cannot be estimated in Excel The initial estimates in a more-advanced package are coefficient estimates not marginal effects We should convert the coefficient estimates to marginal effects before interpreting them
10
12-10 The Logit Model Example:
11
12-11 The Logit Model
12
12-12 The Probit Model The probit model is a popular specification for estimating population regression models with binary dependent variables. Notes: The Probit model cannot be estimated in Excel The initial estimates in a more-advanced package are coefficient estimates not marginal effects We should convert the coefficient estimates to marginal effects before interpreting them Probit Function
13
12-13 The Probit Model Example:
14
12-14 The Probit Model
15
12-15 Comparison of the Models
16
12-16 Categorical Dependent Variables A categorical dependent variable is a dependent variable that takes on a limited, and usually fixed, number of possible values.
17
12-17 Categorical Dependent Variables Example: Q:Do you have enough food to eat? 2=Always 1 =Sometimes 0 = Never Data: 17,723 individuals drawn from the National Health and Nutrition Examination Survey
18
12-18 The Multinomial Logit The multinomial logit is a regression model that generalizes logistic regression to allow more than two discrete outcomes for the dependent variable. Note: The Multinomial Logit model cannot be estimated in Excel The initial estimates in a more-advanced package are coefficient estimates not marginal effects We should convert the coefficient estimates to marginal effects before interpreting them
19
12-19 The Multinomial Logit For example, Do not interpret these directly Convert them to marginal effects first
20
12-20 The Multinomial Logit
21
12-21 The Multinomial Logit
22
12-22 The Ordered Probit The ordered probit a regression model that generalizes the probit model to allow more than two discrete outcomes for the dependent variable. To be the appropriate model the outcomes of the dependent variable must be ordinal. Note: The Ordered Probit model cannot be estimated in Excel The initial estimates in a more-advanced package are coefficient estimates not marginal effects We should convert the coefficient estimates to marginal effects before interpreting them
23
12-23 The Ordered Probit For example, Do not interpret these directly Convert them to marginal effects first
24
12-24 The Ordered Probit
25
12-25 The Ordered Probit
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.