Python Lesson 4: Logistic Regression for a Binary Response Variable, pt. 2

Loading...

Do curso por Wesleyan University

Regression Modeling in Practice

217 classificações

Wesleyan University

Regression Modeling in Practice

217 classificações

Curso 3 de 5 em Specialization Análise e Interpretação de Dados

This course focuses on one of the most important tools in your data analysis arsenal: regression analysis. Using either SAS or Python, you will begin with linear regression and then learn how to adapt when two variables do not present a clear linear relationship. You will examine multiple predictors of your outcome and be able to identify confounding variables, which can tell a more compelling story about your results. You will learn the assumptions underlying regression analysis, how to interpret regression coefficients, and how to use regression diagnostic plots and other tools to evaluate the quality of your regression model. Throughout the course, you will share with others the regression models you have developed and the stories they tell you.

Na lição

Logistic Regression

In this session, we will discuss some things that you should keep in mind as you continue to use data analysis in the future. We will also teach also you how to test a categorical explanatory variable with more than two categories in a multiple regression analysis. Finally, we introduce you to logistic regression analysis for a binary response variable with multiple explanatory variables. Logistic regression is simply another form of the linear regression model, so the basic idea is the same as a multiple regression analysis. But, unlike the multiple regression model, the logistic regression model is designed to test binary response variables. You will gain experience testing and interpreting a logistic regression model, including using odds ratios and confidence intervals to determine the magnitude of the association between your explanatory variables and response variable. You can use the same explanatory variables that you used to test your multiple regression model with a quantitative outcome, but your response variable needs to be binary (categorical with 2 categories). If you have a quantitative response variable, you will have to bin it into 2 categories. Alternatively, you can choose a different binary response variable from your data set that you can use to test a logistic regression model. If you have a categorical response variable with more than two categories, you will need to collapse it into two categories.

Python Lesson 3: Logistic Regression for a Binary Response Variable, pt. 17:27

Python Lesson 4: Logistic Regression for a Binary Response Variable, pt. 23:19

Conheça os instrutores

Jen Rose
Research Professor
Psychology
Lisa Dierker
Professor
Psychology

But what happens if we control for major depression?

As you can see, both social phobia and major depression are independently

associated with the likelihood of having nicotine dependence.

Given that both social phobia and major depression are positively associated with

a likelihood of being nicotine dependent.

In our predictor or splinter variables are both binary,

we can interpret the odds ratio in the following way.

Young adult daily smokers, the sample population with social phobia

are 2.3 times more likely to have nicotine dependence than young adult

daily smokers without social phobia after controlling for major depression.

Also, daily smokers with major depression are 3.7 times more

likely to have nicotine dependence than daily smokers without depression,

after controlling for the presence of social phobia.

Because the confidence intervals on our odds ratios overlap,

we can not say that major depression is more strongly associated with nicotine

dependence than the social phobia.

For the population of young adult daily smokers,

we can say that those with social phobia are anywhere between 1.2 to 4.6 times

more likely to have nicotine dependence than those without social phobia.

And those with major depression are between 2.7 and 5.0 times more likely to

have nicotine dependence than those without major depression.

Both of these estimates are calculated after accounting for

the alternate disorder.

As with multiple regression, when using logistic regression,

we can continue to add variables to our model in order to evaluate multiple

predictors of our binary categorical response variable.

Presence or absence of nicotine dependence symptoms.

Another example of confounding occurs when a logistic progression model is run to

test the association between panic disorder as the explanatory variable and

nicotine dependence, the response variable.

Panic disorder is an anxiety disorder characterized by recurring panic attacks.

Here we see a significant association, and note that young adult daily smokers with

panic disorder in our sample are 2.1 times more likely to have

nicotine dependence than young adult daily smokers without panic disorder.

However, when we add major depression to the model,

panic disorder is no longer significantly associated with nicotine dependence.

Here we have an example of confounding.

We would say that major depression confounds the relationship between panic

disorder and nicotine dependence because the P-Value for panic disorder

is no longer significant when major depression is included in the model.

Further, because panic disorder is no longer associated with nicotine

dependence, we would not interpret the corresponding odds ratio, but

would interpret the significant odds ratio between major depression and

nicotine dependence.

That is, young adult smokers with major depression are 3.6 times more likely

to have nicotine dependence than young adult smokers without major depression

after controlling for panic disorder.

>> By now you should be feeling a little more comfortable with

the idea of generating a regression model when your outcome variable is binary.

Remember to always code your outcome variable so

0 means no outcome, and a 1 means that an outcome occurred.

This is true whether your outcome is positive, such as graduating from college,

or negative, such as developing nicotine dependence.

Explore nosso catálogo

Registre-se gratuitamente e obtenha recomendações, atualizações e ofertas personalizadas.

O Coursera proporciona acesso universal à melhor educação do mundo fazendo parcerias com as melhores universidades e organizações para oferecer cursos on-line.

© 2018 Coursera Inc. Todos os direitos reservados.

Baixar na App Store

Baixar no Google Play