STATA Tutorials: Binary Logistic Regression

welcome to the LSC methodology institute state a tutorial series sponsored by the LLC and well fund in this video we will show how to run a binary logistic regression hello again as with the other tutorial videos we'll be using the data set mi underscores data tutorial DTA in this video we'll be showing you how to carry out a binary logistic regression analysis in other words the response variable is a binary variable a variable with two categories the variable we'll be using as a response variable is this variable here called petition which asked respondents whether or not they've ever signed a petition for our explanatory variables will be using gender a binary variable which asks the respondents gender the continuous variable age which asks the respondents age in years and finally this binary variable here called vote which asks responds whether or not they intend to vote in the next general election so to carry out a logistic regression analysis in Stata it's quite straightforward you type in the command low jet and then first select your response variable petition and then select your explanatory variables gender age and vote and then click enter and wait for Stata to give you output right so now instead of getting coefficients if we wanted to get odds ratios instead of using the command logit we would use the command logistic and then the rest of the command is the same as before where we select our response variable petition and then we click on our explanatory variables gender age and vote and click enter and wait for this data output so now we have some output from this logistic regression analysis what we'll do next is explain what some of these numbers mean and here we have the how to put for the binary logistic regression we have two tables the first table was generated by the command logit and the second table was generated by the command logistic the difference between the two tables is that in the first table we have coefficients and in the second table we have odds ratio if you look at the coefficients we can see that the coefficients are negative for gender and age and for vote the coefficient is positive we are testing if these coefficients are equal to zero and we have the Z test here and the p-value associated to the test we conclude that those people who intend to vote at the next general election are more likely to f sign of a petition in the past we have the confidence intervals for the coefficient of vote here and in the second table we have to authorize you and you are testing the same hypothesis we are testing that the coefficients are equal to 0 or the odds ratio are equal to 1 and we get the same results because we are testing the same thing only the coefficient or the odds ratio for vote really significant and the authorised you are here you get to authorize you by taking the exponential of the coefficients and odds ratio for vote is six point seven which means that the odds of those people who intend to vote at the next general election are six point seven times higher than the odds of the group who intend not to vote at the next general election and we have the 95% confidence interval for odds ratio here goodbye

Channel: Department of Methodology LSE

Views: 93,986

Rating: 4.630137 out of 5

Keywords: STATA, STATA Tutorial, software tutorial, data, Logistic Regression, data analysis, STATA analysis, Methodology Institute, Binary Logistic Regression, statistical software, STATA Linear Regression, Binary Regression, Binary Logistic

Id: 0C_Hlh_jNq8

Length: 3min 50sec (230 seconds)

Published: Fri Jun 29 2012