Confirmatory Factor Analysis Using Stata (Part 1)

[Music] this video is designed to demonstrate how to use data 14 to conduct a confirmatory factor analysis for this example we will use data from the classic holes in German Schweinfurt research study that collected data on 26 psychological tests from seventh and eighth grade children the model that we're going to test is this model here to determine if we can fit a spatial and a verbal construct to three visual spatial sub tests or tasks that measure how well students are able to use their hands on kinds of skills to learn hands on visual spatial skills and then we'll also test to see if these three verbal sub tests will then work together to measure a verbal ability construct the data currently that I have is in SPSS and in order to use Stata we have to save the SPSS data file as a Stata data file so any of your statistical programs that are have that hold your data will need to be converted in some manner to a Stata data file it's very easy in SPSS we can just go to save as and then you can see I've already saved at once but I'll go ahead and save it again and I can scroll down to version 14 a Stata and then I can rename it's a data file and I'll save that so now that I have the data file saved I can go ahead and close SPSS down and then I'm going to open up my folder that I have my static data and you can see it saved this holzinger and swine for data with a dot d ta suffix and that's or extension and that's the Stata how it labels how Stata labels its files so one of the things we can do is I could double click on this and open Stata or I can go down here and open the program and then what I'm going to do is go to file open and I'm gonna open that Stata data file that we just saved from SPSS now you notice like in you'll notice that what Stata does not do is it does not automatically open the data into a spreadsheet like SPSS does and so if you want to view that data you'll have to go to the data editor and then go to browse and then you can see what the data set looks like so we're gonna close that out and then we're going to begin our process of our confirmatory factor analysis so we go to statistics and we'll go to SEM structural equation modeling and then model building an estimation and what you notice here is we get this what's called a graphical user interface and so this will allow us to draw our model and what we're going to do is we'll look at that model again and I'm just gonna put it over here so we can kind of use it as a guide and so to form our first measurement construct we'll click on this little icon to add the measurement component and then we'll just click right about there and so we're gonna name this spatial and then what we're going to do is add these three sub tests and so we'll come down here and we'll add those measurement variables visual perceptual cubes and Long's inch wands and by the way is a paper folding activity and we can put our measurement variables to the left of our construct to the right up or down and in this case I'm going to keep it to the left so I'm going to go okay and you can see how it's formed that graphic now if I want to manipulate this at all I've got to click on the arrow key and then I can pull that down so we can see it so now I'm going to go ahead and form the the verbal construct and I'm gonna go to my measurement component again and then I'm gonna call this verbal and then I'll add my sub tests I'm gonna add paragraph completion sentence reading and word meaning then I'm gonna go okay I want to keep my measurement variables my indicator variables to the left again and go okay and again if I want to align these a little better I can click on the arrow up above and do that the other thing we need to do is you'll notice this has a covariance or a correlation indicator between spatial and verbal and so we click on this double sided arrow to add that covariance and so now we have our model formed the other thing I wanted to demonstrate is if you want to if you have a more complex model and want to make your drawing canvas law largerr you can't you can do it using this procedure you can make it wider and you can make it longer if you like now to estimate this model based on the data that we have in Stata we go to estimate and we have some different choices but we typically use maximum likelihood and then for reporting what I like to do is get the standardized coefficients and values and so that I'm gonna I'm gonna close this pitcher down and then I'm gonna go okay and so automatically what Stata does is it outputs your estimates okay and you'll also notice that if we toggle back to the window those estimates are also represented graphically now you can save this diagram so you can go save as and we can label it and so you can save it in your folder and what it does is then it saves it as a let's see we're going to put it it saves it as @ s T SEM file and if you click on that and open it again it'll maintain those same estimates for you and so if we go back to the output page the results one of the things we want to be able to look at in terms of model fit are the goodness of fit statistics so we can go up to statistics we can go back to SEM and then goodness of fit and overall and in this case we can see the goodness of fit is highlighted and you can get the chi-square so you can get single goodness of fit indices but I typically click all and then when I run this you can see that we get the chi-square statistics and what most people are interested in is the root mean square error of approximation and you can see for this particular analysis this point only two which is a marginal fit and we can go back and respess off' i the model to improve that fit but that would be for another that would be for another time in in doing that analysis

Channel: Arthur Bangert

Views: 13,957

Rating: 4.9375 out of 5

Keywords:

Id: nfhHkSfacTI

Length: 9min 37sec (577 seconds)

Published: Wed Feb 07 2018