Intro to using AMOS with regression and path analysis

[@mikecrowson2462]

Hello everyone: I have a much newer (and higher quality) video (Introduction to path analysis with manifest variables using AMOS Oct 2020) you can go to here: kzbin.info/www/bejne/b6XQZYynibiAbK8 This video covers the same (for the most part) topics, with some updated info. FYI, underneath the video description is a link to a Powerpoint (with citations) you can download, as well as an SPSS data file. I hope you will consider visiting!

[@hereisJH]

thank you so much mike,,,, your video helped a lot...

[@tetlleyplus]

Is it my PC, or the video and audio are not synchronized?

[@tetlleyplus]

Michael Crowson, do you have a version in which de audio is sync? The video is several seconds ahead of the audio.

[@sue9679]

Hi Mike, thanks for your quality video! It was very helpful:) I have one question regarding model fit. If I am conducting a path analysis with predictor variables (not latent variables) that I assessed, is it important to look at & report model fit?

[@xijialiu5740]

thanks for your nice video. I have a question about the intercept term in the regression model. It seems that AMOS includes intercept term by default, then how should I remove the intercept term? I am just wondering how to control the intercept term in AMOS. Thanks in advance

[@xijialiu5740]

I worried about the intercept term before because I was not sure how to remove the constant term in a factor analysis model (factor should be a weighted sum of variables without intercept term) Now, I probably figure it out. Amos provides the function of estimating the model parameters from the mean vector and covariance matrix, so the regression model contains the intercept term by default. In other words, one can not fit a regression model without constant terms in Amos. However, this rule only happens in the mean structure part, but not in the covariance structure part. So, there is no intercept term by default in a CFA model, i.e. the latent variable is a weighted sum of observed variables without a constant term. The same rule is applied in a more general structural equation model.

[@ignaciojaviersalamancagara2091]

Hola, muy bueno tu video, podrías agregar la opción de subtitulos en español por favor, es muy importante para mi intender bien lo que explicas en el video, gracias.

[@Xylomag]

Hi Michael, I'm currently learning about SEM and finding your videos really helpful! I just wanted to ask a few Qs about path analysis if that's ok with you. If I wanted to see the relationship between my two predictor variables would I just look at the correlation output? And to confirm if the relationship between the two predictors is reliably different from 0 would I just look at my covariance output (e.g is significance and has covariance of .79). Additionally, looking at the variances table, how would I see which predictor is best contributing to the model, I see that it has significances but not sure how to interpret the estimates (are these their amount of variance explained?) thank you and hope these aren't too many Qs!

[@mikecrowson2462]

Hi Sirin, if you asked for standardized estimates (by clicking on Standardized Estimates under the Output tab when specifying your analysis properties), then you can get Pearson's correlation between the exogenous variables. From there, you can report on the significance of the correlation by referencing the test associated with the covariance (note: Pearson's correlation is simply the covariance between two standardized variables). If you are seeking to make a comparative judgment concerning (direct) predictors of a specific endogenous variable, then you can look at the standardized partial regression coefficients found in the output (again, you'll have to click on Standardized Estimates under the Output tab). If you have several predictors of a given exogenous variable, you can rank according to their absolute values. The predictor associated with the largest absolute value would be the strongest predictor of an outcome variable. The next largest absolute value would indicate the next strongest predictors, etc.

[@mamathad2601]

in this video regression model file is not given. can you please provide

[@mikecrowson2462]

HI Mamatha, you should be able to obtain the datasets for this video by following the links provided in the video description. Also, just an FYI, I have a much newer video on AMOS basics here: kzbin.info/www/bejne/hpyrpZKmhM-LmK8 . This video (from Jan 2021) contains a general walkthrough, including of a regression model. Another video (from October 2020) can be found here (kzbin.info/www/bejne/b6XQZYynibiAbK8) on path analysis with AMOS. The data for each of those videos and supplemental files can be downloaded at the links in the video description. Cheers!

[@mamathad2601]

@@mikecrowson2462 Thank you so much for responding. But I didn't understand what you did with customised estimate. I thought you were selecting some file And that file I could not found. But later I understood you were saving the model before running. Thank you

[@emzalmighty]

Audio is not in sync.

[@donnaharrilal9936]

the video is not matching your lecture unfortunately