## Watch “H-PH207X-FA12-L11-9_100” on YouTube

Calculate propesity score

Probability of each individual to be in the exposed and non-exposed group

Match the probabilty (propensity score) for each individual in the exposed and non exposed group

## Binary dependent variable – Logistic regression

This will show

1. the analysis for binary outcome
2. why linear is not good
3. the stata command
4. the predicted probability using ‘margins’
5. measure firness using fitstat

Note to Logistic Regression

## Measuring goodness of fit in logistic regression Using R

SOURCE : http://www.medicine.mcgill.ca/epidemiology/joseph/courses/epib-621/logfit.pdf

This is a good summary on the measure of the goodness of fit in logistic regression. The data can be searched online. It was based on the icu data available from the Hosmer-Lemeshow logistic regression book

The analysis in R.

The discussion is on:

1. Chi-square goodness of fit tests and deviance
2. Hosmer-Lemeshow tests
3. Classification tables
4. ROC curves
5. Logistic regression R2
6. Model validation via an outside data set or by splitting a data set

PDF can be downloaded here: logfit

## Binary outcome data with Generalized Linear Model

The author uses the glm function to analyze binary outcome data. If you use STATA, you maybe familiar with logit or logistic function because glm command in stata is rarely used.

http://ecology.msu.montana.edu/labdsv/R/labs/lab4/lab4.html