1 G89.2229 Lect 11M Binary outcomes in psychology Can Binary Outcomes Be Studied Using OLS Multiple Regression? Transforming the binary outcome Logistic.

Slides:



Advertisements
Similar presentations
Statistical Analysis SC504/HS927 Spring Term 2008
Advertisements

Qualitative predictor variables
Brief introduction on Logistic Regression
Regression analysis Linear regression Logistic regression.
1 Multiple Regression Response, Y (numerical) Explanatory variables, X 1, X 2, …X k (numerical) New explanatory variables can be created from existing.
Regression With Categorical Variables. Overview Regression with Categorical Predictors Logistic Regression.
April 25 Exam April 27 (bring calculator with exp) Cox-Regression
Logistic Regression Multivariate Analysis. What is a log and an exponent? Log is the power to which a base of 10 must be raised to produce a given number.
N-way ANOVA. 3-way ANOVA 2 H 0 : The mean respiratory rate is the same for all species H 0 : The mean respiratory rate is the same for all temperatures.
1 BA 275 Quantitative Business Methods Residual Analysis Multiple Linear Regression Adjusted R-squared Prediction Dummy Variables Agenda.
January 6, afternoon session 1 Statistics Micro Mini Multiple Regression January 5-9, 2008 Beth Ayers.
EPI 809/Spring Multiple Logistic Regression.
Nemours Biomedical Research Statistics April 23, 2009 Tim Bunnell, Ph.D. & Jobayer Hossain, Ph.D. Nemours Bioinformatics Core Facility.
(Correlation and) (Multiple) Regression Friday 5 th March (and Logistic Regression too!)
An Introduction to Logistic Regression
Generalized Linear Models
C. Logit model, logistic regression, and log-linear model A comparison.
Logistic regression for binary response variables.
Unit 4c: Taxonomies of Logistic Regression Models © Andrew Ho, Harvard Graduate School of EducationUnit 4c – Slide 1
1 1 Slide © 2008 Thomson South-Western. All Rights Reserved Slides by JOHN LOUCKS & Updated by SPIROS VELIANITIS.
Quantitative Business Analysis for Decision Making Multiple Linear RegressionAnalysis.
Unit 4c: Taxonomies of Logistic Regression Models © Andrew Ho, Harvard Graduate School of EducationUnit 4c – Slide 1
MODELS OF QUALITATIVE CHOICE by Bambang Juanda.  Models in which the dependent variable involves two ore more qualitative choices.  Valuable for the.
Inference for regression - Simple linear regression
1 G Lect 11W Logistic Regression Review Maximum Likelihood Estimates Probit Regression and Example Model Fit G Multiple Regression Week 11.
G Lecture 121 Analysis of Time to Event Survival Analysis Language Example of time to high anxiety Discrete survival analysis through logistic regression.
1 1 Slide © 2007 Thomson South-Western. All Rights Reserved Chapter 13 Multiple Regression n Multiple Regression Model n Least Squares Method n Multiple.
Logit model, logistic regression, and log-linear model A comparison.
ALISON BOWLING THE GENERAL LINEAR MODEL. ALTERNATIVE EXPRESSION OF THE MODEL.
April 6 Logistic Regression –Estimating probability based on logistic model –Testing differences among multiple groups –Assumptions for model.
Business Intelligence and Decision Modeling Week 11 Predictive Modeling (2) Logistic Regression.
1 G Lect 9M Example Nominal (Categorical) Variables as Explanatory Factors Coding of Nominal Explanatory Variables G Multiple Regression.
1 G Lect 7M Statistical power for regression Statistical interaction G Multiple Regression Week 7 (Monday)
University of Warwick, Department of Sociology, 2014/15 SO 201: SSAASS (Surveys and Statistics) (Richard Lampard) Week 7 Logistic Regression I.
Multiple Regression and Model Building Chapter 15 Copyright © 2014 by The McGraw-Hill Companies, Inc. All rights reserved.McGraw-Hill/Irwin.
When and why to use Logistic Regression?  The response variable has to be binary or ordinal.  Predictors can be continuous, discrete, or combinations.
April 4 Logistic Regression –Lee Chapter 9 –Cody and Smith 9:F.
Assessing Binary Outcomes: Logistic Regression Peter T. Donnan Professor of Epidemiology and Biostatistics Statistics for Health Research.
Regression. Types of Linear Regression Model Ordinary Least Square Model (OLS) –Minimize the residuals about the regression linear –Most commonly used.
Copyright © 2006 The McGraw-Hill Companies, Inc. All rights reserved. McGraw-Hill/Irwin Dummy Variable Regression Models chapter ten.
Business Intelligence and Decision Modeling
Lecture 12: Cox Proportional Hazards Model
1 Multivariable Modeling. 2 nAdjustment by statistical model for the relationships of predictors to the outcome. nRepresents the frequency or magnitude.
Warsaw Summer School 2015, OSU Study Abroad Program Advanced Topics: Interaction Logistic Regression.
28. Multiple regression The Practice of Statistics in the Life Sciences Second Edition.
Multiple Logistic Regression STAT E-150 Statistical Methods.
Multiple Regression  Similar to simple regression, but with more than one independent variable R 2 has same interpretation R 2 has same interpretation.
1 1 Slide © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole.
Logistic Regression. Linear regression – numerical response Logistic regression – binary categorical response eg. has the disease, or unaffected by the.
Heart Disease Example Male residents age Two models examined A) independence 1)logit(╥) = α B) linear logit 1)logit(╥) = α + βx¡
Logistic regression (when you have a binary response variable)
1 Say good things, think good thoughts, and do good deeds.
Introduction to Multiple Regression Lecture 11. The Multiple Regression Model Idea: Examine the linear relationship between 1 dependent (Y) & 2 or more.
1 Introduction to Modeling Beyond the Basics (Chapter 7)
Probability and odds Suppose we a frequency distribution for the variable “TB status” The probability of an individual having TB is frequencyRelative.
1 G Lect 3M Regression line review Estimating regression coefficients from moments Marginal variance Two predictors: Example 1 Multiple regression.
Logistic Regression Hal Whitehead BIOL4062/5062.
Roger B. Hammer Assistant Professor Department of Sociology Oregon State University Conducting Social Research Logistic Regression Categorical Data Analysis.
Logistic Regression and Odds Ratios Psych DeShon.
Logistic Regression Logistic Regression - Binary Response variable and numeric and/or categorical explanatory variable(s) –Goal: Model the probability.
LOGISTIC REGRESSION. Purpose  Logistical regression is regularly used when there are only two categories of the dependent variable and there is a mixture.
1 BUSI 6220 By Dr. Nick Evangelopoulos, © 2012 Brief overview of Linear Regression Models (Pre-MBA level)
BINARY LOGISTIC REGRESSION
Logistic Regression APKC – STATS AFAC (2016).
Applied Biostatistics: Lecture 2
Generalized Linear Models
Multiple logistic regression
THE LOGIT AND PROBIT MODELS
Prepared by Lee Revere and John Large
Introduction to Logistic Regression
Presentation transcript:

1 G Lect 11M Binary outcomes in psychology Can Binary Outcomes Be Studied Using OLS Multiple Regression? Transforming the binary outcome Logistic Regression G Multiple Regression Week 11 (Monday)

2 G Lect 11M Binary outcomes in psychology Psychologists tend to study quantitative processes such as reaction time, strength of attitudes, achievement scores, and severity of depression In some cases the outcome is binary or categorical »Decisions (e.g. to sit next to certain person) »Passing some threshold to a new qualitative state (such as major depression) »Death, illness, stressful event »Promotion, retirement, selection

3 G Lect 11M Example: Diagnosis of Major Depression in Puerto Rican Adolescents Adolescents were assigned diagnoses on the basis of either their or parents’ reports The binary outcome makes a bad plot.

4 G Lect 11M Binary Outcomes and OLS Multiple Regression Why not regress a (0,1) binary outcome Y on X? »Y=B 0 +B 1 X 1 +B 2 X 2 +e »Interpret E(Y|X) as Prob(Y=1|X) E(Y|X)= B 0 +B 1 X 1 +B 2 X 2 Puerto Rican Example »Y is major depression in past year (0 absent, 1 present) »X 1 is centered age »X 2 dummy code for female gender »Perhaps the risk of depression goes up linearly with each year

5 G Lect 11M Numerical Results The probability of depression among male 15 year olds appears to be.026 in the community sample. The probability of depression in the clinic sample appears to be.125 higher than in the community sample. For each year of age, the probability seems to go up.013. The probability for females is.031 higher than males. The probability for 10 year old community males is negative.

6 G Lect 11M Formal Objections to OLS Analysis Linear formula for E(Y|X) may lead to values that exceed the logical (0,1) interval. Var(Y|X) is not constant »If binary variable is thought to be a binomial random variable, E(Y|X)=p X, Var(Y|X)= p X (1- p X ) »We expect (0,1) variation around region.4< p X <..6 »We don't expect much variation when p X approaches 0 or 1 (e.g. p X =.01 suggests Y=1 will almost never be seen.) »OLS is not efficient

7 G Lect 11M Residual distribution from depression example This distribution is not normal. Variation of residuals depend on probability of outcome.

8 G Lect 11M Transforming the binary outcome We need a transformation of E(Y|X) that is unbounded and that is easy to interpret. »The odds, p X /(1- p X ), is used by gaming fans. "The odds are 2 to 1 that our team will win" odds are unbounded in the positive direction, but still bounded by 0 »p X =.5 gives odds of 1.0 »p X =>  gives odds=  »p X =>0  gives odds=0 Taking the ln of the odds makes it unbounded in both directions

9 G Lect 11M Ln odds:  The transformation ln[p X /(1- p X )] gives a metric that is centered around p X =.5 and ranges from  to . »ln[p X /(1- p X )] =  X »When p X =.5,  X = 0.

10 G Lect 11M Logistic Regression The ln(odds)=  is called a logit. A regression model with  as the outcome is called logistic regression »  = B 0 + B 1 X The parameters of logistic regression have a natural interpretation when their antilog (exp) is taken »The constant is simply the odds of the outcome when X=0 »The slope becomes an odds ratio

11 G Lect 11M Example with Logistic Regression Constant indicates the odds of depression among 15 yr male community is.026 »Probability is similar,.025 The odds increase by a factor of for each year of age. The odds increase by a factor of 1.65 for females. The odds increase for clinic patients

12 G Lect 11M Plot of Fitted Probabilities with Age Females in the clinic sample have substantially higher risk when they are older. The interaction of sex by age is not significant.

13 G Lect 11M Details of Interpreting Logistic Coefficients Consider a simple model with one X dummy (0,1) variable »  = B 0 + B 1 X when X=0 »E(  X=0) =  0 = B 0 B 0 is the ln odds in the reference group exp(B 0 ) is the odds of Y=1 in group 0 when X=1 »E(  X=1) =  1 = B 0 + B 1

14 G Lect 11M Transforming Logits for Interpretation exp(   ) is the odds of Y=1 in group 0 exp(   ) is the odds of Y=1 in group 1 »B 1 is the odds ratio.