Stata ordered logistic regression

hi there and welcome to 272 analytics comms tutorial on how to use ordered logistic regression in Stata in order to do so let's go ahead and load the full-auto dataset it's a pretty comprehensive data set of the qualities of cars in Stata and the variables we'd like to check on today are rep 77 which is repair record and location of manufacture which is this called foreign in here so we have domestic and foreign as our two possibilities in order to do the regression the ordered logistic regression let's start with a chi squared which is what Stata documentation also recommends we'll go ahead and do that and we see that it's significant so there does seem to be some connection between the origin of the car in terms of manufacturing and its repair record now given that and given that the repair record is ordinal in nature we have the scope to use an ordered logistic regression which is simply ole oujet followed by or your dependent variable which is ordered here and as you can see rep 77 goes from for fair average good so it's ordinal and a foreign as your predictor variable just type enter and we see that it is in fact significant and we want to know how to interpret it and this is where it gets a little bit more complicated so I'm going to pull up notepad so you can actually see you know how we can take these values here and make sense of them in terms of odds now cut one is basically for poor so it's it's it's the odds that a foreign or a domestic car resulted in a poor repair record that's what cut one is and you know you can just check these cuts versus the actual ordinal values of the repair record and and do that analysis for any particular slice what I'm going to demonstrate is again building on the the standard documentation how you get to the odds for a poor repair record for foreign and domestic cars based on euro low jet procedure so here's the notepad here and let's start by noting that foreign has a coefficient which I've gone ahead and represented as 1.46 and since it has this coefficient domestic as a coefficient of zero and in a moment we're gonna see how and why that's important now here you can also see the cut point which is again negative two point seventy seven just about so the first thing we have to do is we have to get to U of J okay and the formula is one point four six plus u FJ is less than or equal to negative two point seven seven so then we can set u of J less than or equal to negative four point two three and now we use the logistic distribution formula which is one divided by one plus e raised to the power of four point two three and this becomes positive here by the way not negative and when we carry out the calculation we see that the odds are point zero one four percent for foreign cars what does that mean well there's a one point four percent chance that a foreign car is going to get a poor repair record so that's pretty informative now we've followed the same procedure here with the difference being that there is no coefficient for domestic okay think of it almost like it's a dummy variable in that sense because of that you don't have to insert this number into this formula it's just UF J is less than or equal to negative two point seven seven and so the formula is this it's one divided by one plus e raised to the two point seven seven and when we calculate that out it's zero point zero five nine so there is a roughly six percent chance that a domestic car is going to have a poor repair record so that's the way you go from you know a couple of things you're going to start from a chi-square table as Stata recommends and then you move on to the old legit procedure and you generate these cut points and basically all I've done is I've walked you through what the standard documentation also illustrates as you know the way of applying ole logit to these data i hope it was helpful also to see the text to show the formula to go ahead and calculate the odds from the cut points so that's something you can manually do with each of these cut points I hope this tutorial was helpful to you and I would like to invite you to visit 272 analytics comm for access to all our free statistics tutorials in Stata SPSS are eviews and Minitab here at 272 analytics comm we provide data consulting primarily to graduate students therefore we work very closely with you in order to perfect your chapter 3 and chapter 4 that means helping you design surveys getting your data input assisting with fashioning appropriate research questions and hypotheses getting your data extracting them transforming them cleaning them putting them through analysis interpreting them explaining them to you so that at the end of the day you know exactly what story your data tell why they matter what they mean in a manner that lets you write a perfect chapter 4 of following a perfect chapter 3 and lets you defend your dissertation or thesis with complete confidence we provide ethical consulting it's not a writing service so you will be responsible for taking our blueprint our assistance our consulting and transforming them into an appropriate academic project for yourself I'd also like to remind you that we provide the same services to undergraduate students who are working with quantitatively oriented assignments thank you so much for listening and have a great day

Channel: 272analytics Videos

Views: 20,762

Rating: 4.8048782 out of 5

Keywords: Stata, ordered logistic regression, logistic regression, statistics, chi-square

Id: -w3tEguP5-w

Length: 5min 59sec (359 seconds)

Published: Sun Dec 03 2017