Moderated multiple regression using Hayes' Process Macro v3.3 with SPSS (July 2019)
HTML-код
- Опубликовано: 22 авг 2024
- In this video, I provide a walk-through of moderated multiple regression using Andrew Hayes' Process macro with SPSS. I go over four examples and discuss interpretation of output. I also discuss mean centering, simple slopes tests, etc.
The Powerpoint presentation I reference in the video can be downloaded (I encourage this!) at:
drive.google.c...
A copy of the data can be downloaded here:
drive.google.c...
To learn how to download the Process macro (as of 2023), go here:
• Downloading and instal...
For more videos on regression analysis, download a document containing a list of links here: drive.google.c...
NEW 2020 VIDEO ON USING THE MULTICATEGORICAL OPTION IN PROCESS WHEN PERFORMING MODERATED REGRESSION: • Moderated multiple reg...
Thanks for watching!
this is super super helpful! trying to learn about moderation analysis for my dissertation!!
Glad to be of help! Good luck to you!
Thank you very much...your video is the best perceptible than other slope test videos...
You're welcome Ibrahim. Best wishes!
@@mikecrowson2462 How or which model can I use if I am using three(3) moderating variable
For example, what are the moderating effects of A, B, and C on the relationship between X on the growth of Y in a particular area
I'm very grateful for your presentation sr. It helped me a lot!
You are very welcome, Sefyan! By the way, I have a couple of newer videos on moderated regression using a more recent version of the Process macro. See ruclips.net/video/4gGJCX6rrAo/видео.html and ruclips.net/video/IG-FnhvTud4/видео.html
Best wishes!
This is great Mike! Thank you for sharing your knowledge! I am currently using this to understand moderation which I am using for my MSc research and I find your material extremely helpful!
Nice work! I'm teaching applied multivariate and this video along with your models 4 & 6 video are now in my lesson plan for this semester and going forward.
Dear Ryan, thanks so much for your feedback! I'm so glad you found the videos useful for your students! All the best.
@@mikecrowson2462 what are the differences between model 1 2 3 4 5 6?
@@soehartosoeharto8471 Hi there. There are various differences among these models. Models 1-3 are variations on moderated multiple regression. Model 4 is used to test for mediation. Model 5 involves a mediation model but also testing for moderation of path c. Model 6 is used to test for serial mediation. You can get picture of the different models by following this link: dm.darden.virginia.edu/ResearchMethods/Templates.pdf
Cheers!
@@mikecrowson2462 thanks Prof
very good explanation. thank you so much sir
Super helpful stuff for drawing interaction plots. No need for going online to look for Excel macros to draw plots. Thanks.
SOOO helpful! thank you!
Thank you so much!
Thank you very much. It was an amazing video.
Thank you for your video, this was extremely helpful :)
Great tutorial! It was really helpful for me! Thank you
Thanks for sharing!
You're very welcome. Thanks for visiting, Yulin!
NEW 2020 VIDEO ON USING THE MULTICATEGORICAL OPTION IN PROCESS WHEN PERFORMING MODERATED REGRESSION: ruclips.net/video/IG-FnhvTud4/видео.html
Hi, I have 2 moderators, three IVs and one DV. I want to check the effect of both moderators on the relationship between my IVs and DV simultaneously. Is it Possible to do it in using process macros
Does anyone know of a video that explains moderation output when you have a multicategorical IV with 4 groups??
Thank you so much sir, you explained moderation very excellent manner and also provided PPT, data file also, i hope you can also explain new Hayes' Process Macro v3.4 with SPSS to perform moderated multiple regression
Dear Mike, thank you for your video. Is there a difference between moderated multiple regression (the one you presented in your video) and moderated hierarchical multiple regression??
Hi there. What I presented was moderated multiple regression. I'm not sure what you are asking with respect to 'moderated hierarchical multiple regression'. Are you asking about moderation in the context of hierarchical linear modeling? If that's what you are asking, then they are different procedures (although they both address questions of moderation). In the context of hierarchical linear modeling, you can model predictors of randomly varying regression slopes across clusters/groups. This presumes a multilevel data structure. The video here is addressing moderation as it occurs at a single level. By the way, I do have a more recent video addressing moderated multiple regression here: ruclips.net/video/4gGJCX6rrAo/видео.html
Best wishes!
Hi, sorry I have a question. is this research question correct for analysis of moderator variable through Andrew Hayes ?
To what extent does the self-efficacy moderate the relationship between A and B ?
(Self- efficacy as independent & moderator variable
A as independent variable
B as dependent variable )
What if we have 2.IV and one moderator. One DV. Which model we should choose?
Hi Mike, thanks so much for this helpful video! I'm confused about the direction of the moderation, when the interaction effect is a positive number (b = 2.61), but the conditional effects were negative numbers (the two that were significant anyway - ). Would we still say that grit is a significant negative moderator? Do we go off the conditional effects rather than the total interaction effect to determine the direction of the mediation? Many thanks, Jemma
Hi there. The regression slope for the interaction term can be interpreted as indicating the change in the slope for X on Y as one moves from lower levels of the moderator (W) to higher levels of the moderator in units of one. If it helps, you can say that for every one unit increase on the moderator, the slope for X increases by b3 units (where b3 is the slope for the interaction term). If the slope for the interaction term is positive, it means that the slope for the effect of X on Y is increasing with increasing scores on the moderator. If it is negative, it means that the slope for the effect of X on Y is decreasing with increasing scores on the moderator. The effect of X on Y conditioned on W is expressed as: (b1+b3W)X . If there is no moderator, then the slope for X is simply b1. Since we are modeling the effect as conditional on W, then the slope for X at a given W equals (b1+b3W) . The simple slopes you are referring to as conditional effects at -1sd, at the mean, and +1sd on W. The significance of the simple slopes that are printed out actually say very little about whether the effect is conditional on Y. You can have non-significant slopes or all significant slopes that exhibit enough variability for the interaction term to be significant. The tests of the conditional effects just give us a bit more of an idea of where along the scale on W the effects of X may be significant. I hope that helps!
@@mikecrowson2462 Thank you so much, that was really helpful! :)
what about if you have two variables which are IV with effect on DV and 1 moderator?
Thank you Professor for sharing this 🙂 I had a query, that if my dependent and independent variables are categorical then also we'll be using this models in the same manner? Also can we run these models in Stata?
Thank you for the video sir! Is it possible to plot it without including the main effects to the model and just the interactions? Thank you in advance!
I wonder what if the p value in model summary is significant, the "model section" and in "test of highest order unconditional interaction" are bot nonsignificant, but the output shown the "conditional effect of focal predictor at values of the moderators" and p value in here is significant? Can I say the factor has moderating effect on the relation?
In model 14, if the moderator path coefficient to the dependent variable is not significant while the rest of the path coefficients are significant and the mediator-moderator interaction is also significant, what is the result of the moderator?
hello Mike how you have calculated scores for hope. Sum score of items or Average of items?
I have second order factor as moderator and dont know how to calculate interaction? Should I total and sum the dimensions to generate scores
HI Prof Mike
Wishing you the best of the season.
I am very sorry to ask this question this period. But I think I have no other place to run to. Is it possible to run a moderation analysis with more than two independent variables using Hayes' Process Macro? I have four independent variables, one moderator and one dependent variable. I have been trying to see how I can run this analysis using the Process Macro. Looking forward to hearing from you.
Hi kingsley, I also have the same problem. Do you find any way can solve it? tq
Mike, if I run a model number 1 in the PROCESS 3.3 with a moderator that is multicategorical (you used binary moderator in example 3) then I would expect that I should click on multicategorical tab at start, still enter this (multicategorical) moderator in in the W box and the interpretation of the simple slope (coeff.) for a selected X variable in the model be based (conditional) to the lowest category of my multicategorical moderator. is that right?
Hi Mike, I notice PROCESS doesn't allow for repeated-measures variables. Is there a simple way to conduct a repeat-measures moderation analysis in SPSS? (moderation variable is continuous)
Thanks a lot for sharing this. I'm just wondering if you have a video to test and report moderated multiple logistic regression ( where the DV is logistic) using PROCESS MODEL. I watched your video about plotting logistic regression using Process model.
Hi Maedeh, You're very welcome. And yes, I have a video on using Process with moderated logistic regression. Just go here: ruclips.net/video/Dbwq_or9I0M/видео.html . Thanks for visiting my site!
Thank you very much for this great video! Very helpful. I have a question about the option display in the latest version of SPSS: "Mean center for construction of products" is not available. How can I add it please?
Hi there. This option should be there in the newest version of Process. It looks like Hayes moved in in the most recent version. Here's a video I put together a few weeks ago on the newest version (ruclips.net/video/4gGJCX6rrAo/видео.html). You should see in there. I also have another video you might find interesting on using the multicategorical option (ruclips.net/video/IG-FnhvTud4/видео.html). Cheers!
Hello Mike, thank you very much for your explanations. Would it maybe be possible for you to make a video on model 5 of Hayes extensions? I just cannot find information on how to do the analysis and how to interpret the output and would be super grateful for your help!! :)
Can you please let me know what is ‘F’ in this test?? I mean what does the value of ‘F’ predicts??
Dear Mike
Thank you for sharing such an informative video. I am having one query related to line graph, How you have transferred data from syntax to new SPSS file.
Thanks in advance.
Regards
Hi Sonia, basically if you double click in the Process output file that is generated you will be able to go in and select and copy the syntax. From there, it's a matter of pasting it into a new syntax file and running it. I hope this answers your question! Thank you for visiting!
Thank you so much for your kind reply.
Hi, can you use the Process Macro for repeated measures analysis with moderation and mediation.
I would be eternally grateful if you could tell me how this process will change if we will have multiple independent variables? I am so helpless and my deadline is soon :(
If you have multiple independent variables in your model that are not being modeled as having any interactions with other independent variables, then you can add them into the covariates box. Process Model 1 is only set up to generate the interaction involving X and W in predicting the DV (for a two-way interaction), but you can add in other predictors in the covariates box. You'll get regression coefficients and tests for the other IV's along with the tests of the simple effects of X and W and the interaction term.
@@mikecrowson2462 Thank you so much for your help! So grateful! 😇
Hello sir.. how to apply one-tailed test for the dimensional and non-directional hypothesis. please help
Hi there. The p-values that are printed in the output are given as two-tailed. You can divide the printed p-value to get you the p-value for a one-sided test. Nevertheless, if you are testing a directional hypothesis and you find significance (and with the effect being in the expected direction) using the p-value for the two-sided test, then there's no real need to split the printed p-value in half, as you would be rejecting the null using a more conservative rejection criterion anyways. Hope this helps!
@@mikecrowson2462 Thank you so much, sir, for your spontaneous reply and help.. do you have a reference for this so I could cite it in my paper? It would be very helpful.
@@Mina28000 It's actually just standard hypothesis testing using the t-distribution. It's the same logic that is applied to independent samples t-tests where you are testing directional and non-directional hypotheses. So, I don't have a reference off-hand. Conventionally though, since statistical programs often print out the two-sided p-value, you might see persons reporting the p-value as say, (p
@@mikecrowson2462 thank you :-)
Thank you, what is your profile name in google scholar? I want to follow you there.