Hello. Thank you for the content you create. I would like to ask you, if the variable of the self-reported variable is multicollinear, can I perform a moderation analysis on the amos, interpret the data and report the conclusion?
Hello Dr Crowson, thanks a lot for the guiding videos,, is it posible to edit the estimands file you have provided us for another model (Still MODEL 7)? I am getting a "namespace member declaration" error message in the complie box. my AMOS software does not have a "Define new estimands" function but has "creat a new estimand" would this be the source of the problem?
Thank you so much for this video (and for a lot of others) that has helped a lot. You were mentioning that the Hayes model you used (i. e. a-path being moderated) served as an expample and modeling moderated regression should work for any of the Hayes models. I'm desparately trying to model number 16 where the b-path is moderated. The thing is that I don't get it done. IA-term and so on are all fine. What I don't know ist how to get the covariances/correlations between the predictors and the IA-term (I've been trying only one moderator so far). I simply can't draw them. Where do I go wrong? I'd be really grateful if someone could help me.
Really useful and well-explained, but I would need the same explanation for Model 14 (i.e. moderation of the b-path). Let's see if I can manage to adapt the syntax of the user-defined estimand for this purpose...
Is there a benefit to doing this analysis in AMOS over SPSS? For example, does using AMOS for moderated-mediation provide us with meaningful fit indices?
Hi there. If you use SPSS with the Process macro, you don't get the conventional fit statistics you can obtain if using an Sem program such as Amos. That said, sometimes I think the importance of those fit statistics can be overblown and folks treat them with almost religious conviction. I mainly provided this demo because I've been asked how to do this using Amos. But in terms of flexibility, options and ease of use, Process is way easier than what I demonstrated. I also only demonstrated one model of the large number available through Process. If you wanted to do other process models with the conditional effects and index of moderated mediation for each of those too. My rec is if you are modeling using manifest variables, it's a worthy consideration to use Process. If modeling with latent variables, you'd need the Sem program. Cheers
Hi Mano. I really couldn't give you suggestions on this because the designation of independent variable, mediator, and dependent variable in your model should flow from theory. The specification of proposed causal effects is not a statistical issue but a theoretical one. Cheers!
Thanks for sharing this video and syntax file. However, when I add the mean and standard deviation in user defined estimand file, I am not able compile it successfully. It shows me the syntax errors. The mean of moderator=5.4422, and its standard deviation=0.94936. Please advise if anything else I need to define. Thanks!
Thank you for your video! I calculated the simple slopes to test for the difference in effect of high, medium and low values of my ordinal moderator (latent). My IV is a dummy. Do you have any references to justify the calculation of lowSS=a+c*(mean-sd) (and the other two). I would very much appreciate your answer. Thank you.
Thank you, professor! It was really helpful for me. But after watching the video, I have a question. if a model was constituted of latent variable, then how can I apply this method? Is there any syntax I need to input?
Hi there. Ordinarily I post videos of many procedures using scale scores, so in all likelihood that's how the variables were created for this video (it's been awhile :)
Thank you for the video, we are facing a small problem though. After loading your estimand and clicking on "Edit" button, an error appears stating "hayes model 7 using Amos. SimpleEstimand, is in the wrong format. Can you please help us with this?
Hi there. Well that's strange. Are you copying the text file (under the video description) into the Estimand editor or using the Estimand file (created using the editor; also downloadable from under the video). If you are using the latter, maybe try opening up the text file and copying and pasting it into a new Estimand syntax editor. best wishes!
Hi Sir. the syntax is very helpful. kindly let me know if i can include more than one independent variable to calculate the conditional indirect effect of all the IVs in the single model?
I saw your message here. I would like to know if you are able to run the syntax? I download it, and add mean and deviation, and it show syntax errors when compile it. Appreciate if you could share your experience with me. Thanks!
In the regression model, if the interaction is significant, do we still have to explain the significance of main effect(which is the effect from independent variable)?
Hi Jacob. The interaction effect takes precedence during interpretation as the effect of X --> Y is conditional on Z. The regression slope for the effect of X->Y is simply the effect when the moderator is 0 (for this reason, Hayes, 2016, argues against even referencing the effect as a "main effect" because it is the effect of X on Y at one point along the scale of the moderator), and if the slopes are conditional on Z, then trying to "explain" the effect on Y can't happen without qualifying those effects in your interpretation. You can certainly say that the effect of X on Y when Z=0 is significant or non-significant. And if your X and Z variables had been centered prior to inclusion in the equation, then you could say that it is the effect of X on Y at the mean on Z. But that's really all you can say. (see ruclips.net/video/JZ2fl8exNTE/видео.html) I hope this answers your question!
Hello. Thank you for the content you create. I would like to ask you, if the variable of the self-reported variable is multicollinear, can I perform a moderation analysis on the amos, interpret the data and report the conclusion?
Thanks for your lesson. However, it created an error message 'syntax error' and is not working for AMOS 20. Can you please help?
Hello Dr Crowson, thanks a lot for the guiding videos,, is it posible to edit the estimands file you have provided us for another model (Still MODEL 7)? I am getting a "namespace member declaration" error message in the complie box. my AMOS software does not have a "Define new estimands" function but has "creat a new estimand" would this be the source of the problem?
Thank you Professor. May I know what is the difference if I would like to conduct a mediated moderation model?
Thank you so much for this video (and for a lot of others) that has helped a lot. You were mentioning that the Hayes model you used (i. e. a-path being moderated) served as an expample and modeling moderated regression should work for any of the Hayes models. I'm desparately trying to model number 16 where the b-path is moderated. The thing is that I don't get it done. IA-term and so on are all fine. What I don't know ist how to get the covariances/correlations between the predictors and the IA-term (I've been trying only one moderator so far). I simply can't draw them. Where do I go wrong? I'd be really grateful if someone could help me.
Really useful and well-explained, but I would need the same explanation for Model 14 (i.e. moderation of the b-path). Let's see if I can manage to adapt the syntax of the user-defined estimand for this purpose...
If the moderator is on path b then in estimands would be b+c for direct, lowSS*a for conditional and c*a for Index of moderated mediation?
Thank you so much for this video. How about the moderated mediation for models with a binary moderator?
Wonderful lesson.
Thank you!
How to type estimands in AMOS ? What if we skip it?
Is there a benefit to doing this analysis in AMOS over SPSS? For example, does using AMOS for moderated-mediation provide us with meaningful fit indices?
Hi there. If you use SPSS with the Process macro, you don't get the conventional fit statistics you can obtain if using an Sem program such as Amos. That said, sometimes I think the importance of those fit statistics can be overblown and folks treat them with almost religious conviction. I mainly provided this demo because I've been asked how to do this using Amos. But in terms of flexibility, options and ease of use, Process is way easier than what I demonstrated. I also only demonstrated one model of the large number available through Process. If you wanted to do other process models with the conditional effects and index of moderated mediation for each of those too. My rec is if you are modeling using manifest variables, it's a worthy consideration to use Process. If modeling with latent variables, you'd need the Sem program. Cheers
@@mikecrowson2462 Thank you for the explanation. This was very helpful!
How did you multiply Pime & SVO ? any video that you can share?
Very nice
Thanks Mile!
Thank you very much! Really helpful!
Thank you, professor. Sir link in the description is not working.
Hi, the links should be working now. Cheers!
@@mikecrowson2462 thank you sir
One question please, could you also include SVO in the model ? I mean, as an independent variable predicting BEH: having direct effect on BEH.
There's no reason you couldn't include it as a predictor in the model. Just draw in the direct effect in AMOS. That's all there is to it. Best wishes!
Thankyou
What you think is the mediator in three variables. Sexual harassment. Power relations. And benevolent sexism?
Hi Mano. I really couldn't give you suggestions on this because the designation of independent variable, mediator, and dependent variable in your model should flow from theory. The specification of proposed causal effects is not a statistical issue but a theoretical one. Cheers!
Thank you sir
There are no links in the video description
They are in the video description. Cheers
@@mikecrowson2462 Thanks
Thanks for sharing this video and syntax file. However, when I add the mean and standard deviation in user defined estimand file, I am not able compile it successfully. It shows me the syntax errors. The mean of moderator=5.4422, and its standard deviation=0.94936. Please advise if anything else I need to define. Thanks!
Thank you for your video! I calculated the simple slopes to test for the difference in effect of high, medium and low values of my ordinal moderator (latent). My IV is a dummy. Do you have any references to justify the calculation of lowSS=a+c*(mean-sd) (and the other two). I would very much appreciate your answer. Thank you.
Thank you, professor! It was really helpful for me. But after watching the video, I have a question. if a model was constituted of latent variable, then how can I apply this method? Is there any syntax I need to input?
Did you use averages of the items?
Hi there. Ordinarily I post videos of many procedures using scale scores, so in all likelihood that's how the variables were created for this video (it's been awhile :)
Thank you for the video, we are facing a small problem though. After loading your estimand and clicking on "Edit" button, an error appears stating "hayes model 7 using Amos. SimpleEstimand, is in the wrong format. Can you please help us with this?
Hi there. Well that's strange. Are you copying the text file (under the video description) into the Estimand editor or using the Estimand file (created using the editor; also downloadable from under the video). If you are using the latter, maybe try opening up the text file and copying and pasting it into a new Estimand syntax editor.
best wishes!
Hi Sir. the syntax is very helpful.
kindly let me know if i can include more than one independent variable to calculate the conditional indirect effect of all the IVs in the single model?
I saw your message here. I would like to know if you are able to run the syntax? I download it, and add mean and deviation, and it show syntax errors when compile it. Appreciate if you could share your experience with me. Thanks!
In the regression model, if the interaction is significant, do we still have to explain the significance of main effect(which is the effect from independent variable)?
Hi Jacob. The interaction effect takes precedence during interpretation as the effect of X --> Y is conditional on Z. The regression slope for the effect of X->Y is simply the effect when the moderator is 0 (for this reason, Hayes, 2016, argues against even referencing the effect as a "main effect" because it is the effect of X on Y at one point along the scale of the moderator), and if the slopes are conditional on Z, then trying to "explain" the effect on Y can't happen without qualifying those effects in your interpretation. You can certainly say that the effect of X on Y when Z=0 is significant or non-significant. And if your X and Z variables had been centered prior to inclusion in the equation, then you could say that it is the effect of X on Y at the mean on Z. But that's really all you can say. (see ruclips.net/video/JZ2fl8exNTE/видео.html) I hope this answers your question!
@@mikecrowson2462 Thank you professor! That helps a lot!
Thank you Mike. What if we wanna test a moderator affecting b path?
Thank you so much could please do on Model 58 and Model 16