Excellent video with excellent voiceover and screen demo-best SPSS demo vid i have seen online thus far. I was scratching my head over how to execute LMM precisely(as I read before that it is more complex than traditional rANOVA). Save me tons of time, efforts and frustration! thanks a lot and keep up the good works! :D
I have the following conceptual model: IV: Personality trait (measured on likert scale) Moderator: low trust vs high trust DV: initial price offer in a negotiation ( values can be between 5-15) I have a within-subjects design, so first i assessed the personality of the respondent, then each respondent was asked to input an offer given the scenario (scenario1: low trust towards their negotiation partner). Then in a secod scenario (scenario2: higher trust towards their negotiation partner) they had to input a second offer. I want to measure the effect of personality on the initial offer and how this relationship is moderated by the level of trust. What analysis should I use in SPSS to account for the moderation effect but also for the repeated measures? Do you know which variables should be fixed and which random?
How do you then compare between groups using this? Say if one group of participants had the surgery, and one group didn't (control), where do you add the variable that says which group they were in, and where do you look in the output for the interaction and significant differences?
Hi Scott and everyone. I am doing a repeated measure of stock price and GDP in 5 years. in each year i have 50 number of stock prices , but only 1 annual GDP number. How can i deal with this knind data for a regualr regression analysis? please help me
Hi Scott What if you have another independent variable which is the group type. In other words, you want to look at the time effect as well as the between-group effect (treatment effect). Do you include both independent variables together ?
I have one question. I did an analysis of a 2x2x2x2 factorial anova. The first 3 variables were repeated measures and the last was a between groups analysis. The data in the repeated measures were all reaction speeds. The names of variables were Zoom, Movement, Emotion and Orientation(between). I found two statistical significant interactions between Movement and Emotion, and Zooming and Orientation. Now i want to examine in both interactions which pair of means were statistically different. But I can't do that in the Movement xEmotion interaction with a simple t-test because my dataset has 8 variables and not 4. Also how do I do it in the Zooming x Interaction?
Hi ! Thanks a lot for your explanations. I would like to know how to run a Hausman test on SPSS, in order to choose between a Fixed-effect and random effect model. I have an unbalanced panel data, and I am not sure about what I am doing at all. Thank you !
This video was super helpful! If you have more than one DV, must you run each one individually? Would corrections need to be done in this case? Also, if a particular test spits out multiple statistics that might be correlated, how would we make sure this doesn't interfere with the analysis? Thanks in advance if you have any advice on this!
I am getting this error message: "model cannot be fitted because number of observations is less than or equal to number of model parameters". What does this mean?
Hi! Many thanks for this tutorial. I've got a question for you. Can we use a non-normally distributed dependent variable in these analyses? My variable (anosognosia) is continuous and is kind of normal at time 0 but becomes really skewed in the following 2 waves. Thank you
Jaime P Unfortunately, the Mixed procedure still assumes that the data is reasonably normally distributed. Although, if you have a very large data set, you can relax the assumption of normality.
Dear Scott, Excellent video, it has helped me a lot already. I just have one question: Is there an option to add an across subjects independent variable, for example, splitting the data into male and female patients and seeing if there is also a difference between gender. I know that if the data was parametric, and if there was no missing data, I could use a 'mixed between-within subjects ANOVA' for this, but unfortunately my data is nonparametric and there is missing data, therefore pointing me towards using mixed modelling. Any help on this would be very much appreciated. Regards, Tim
Excellent video with excellent voiceover and screen demo-best SPSS demo vid i have seen online thus far. I was scratching my head over how to execute LMM precisely(as I read before that it is more complex than traditional rANOVA). Save me tons of time, efforts and frustration! thanks a lot and keep up the good works! :D
do u have the video when u introduce other covariates like surgeon or procedure?
Thank you very much!! Could you make an extended video using covariates/random effects/interaction?
Helpful video - could you link to the video where you show how to set up the variable time?
What do you do about the large dfs since they are many rows per participant?
Thanks a lot for this awesome video. But I couldn't understand why you didn't set the ID as a random effect.
I have the following conceptual model:
IV: Personality trait (measured on likert scale)
Moderator: low trust vs high trust
DV: initial price offer in a negotiation ( values can be between 5-15)
I have a within-subjects design, so first i assessed the personality of the respondent, then each respondent was asked to input an offer given the scenario (scenario1: low trust towards their negotiation partner). Then in a secod scenario (scenario2: higher trust towards their negotiation partner) they had to input a second offer.
I want to measure the effect of personality on the initial offer and how this relationship is moderated by the level of trust.
What analysis should I use in SPSS to account for the moderation effect but also for the repeated measures?
Do you know which variables should be fixed and which random?
How do you then compare between groups using this? Say if one group of participants had the surgery, and one group didn't (control), where do you add the variable that says which group they were in, and where do you look in the output for the interaction and significant differences?
Hi! I had the exact same question. Do you happen to have found an answer (three years ago)? Would help a lot!
Hi, can you share which video to watch to see how to create the time variable you referenced? Thanks!
Hi Scott and everyone. I am doing a repeated measure of stock price and GDP in 5 years. in each year i have 50 number of stock prices , but only 1 annual GDP number. How can i deal with this knind data for a regualr regression analysis? please help me
Very helpful video! Saved me a lot of time! Thank you very much!
thank you! at the end of the video you talk about putting in diffrent groups like a diffrent procedure. can you help me with that?
Can you explain us how to introduce covariates and other factors to adjust the model? Thank you very much
Hi Scott
What if you have another independent variable which is the group type. In other words, you want to look at the time effect as well as the between-group effect (treatment effect). Do you include both independent variables together ?
Did you ever figure this out? I'm looking to do the same and am confused! Thanks!
I have one question. I did an analysis of a 2x2x2x2 factorial anova. The first 3 variables were repeated measures and the last was a between groups analysis. The data in the repeated measures were all reaction speeds. The names of variables were Zoom, Movement, Emotion and Orientation(between).
I found two statistical significant interactions between Movement and Emotion, and Zooming and Orientation. Now i want to examine in both interactions which pair of means were statistically different. But I can't do that in the Movement xEmotion interaction with a simple t-test because my dataset has 8 variables and not 4. Also how do I do it in the Zooming x Interaction?
Hi ! Thanks a lot for your explanations. I would like to know how to run a Hausman test on SPSS, in order to choose between a Fixed-effect and random effect model. I have an unbalanced panel data, and I am not sure about what I am doing at all.
Thank you !
This video was super helpful! If you have more than one DV, must you run each one individually? Would corrections need to be done in this case? Also, if a particular test spits out multiple statistics that might be correlated, how would we make sure this doesn't interfere with the analysis? Thanks in advance if you have any advice on this!
also if i have one variable that is repeated measures and another that has between group variance, how do i handle that? thanks in advance
Is this a fixed effect mode only?
Does anyone know which youtube video he is referring to when he says he will explain how he set up "time" for the repeated dialogue box?
Any answer to this question? I would really like to know this
Ah, I see it's explained in this video:
th-cam.com/video/_4pnUa5nmWs/w-d-xo.html
I am getting this error message: "model cannot be fitted because number of observations is less than or equal to number of model parameters". What does this mean?
how we also find R2 change values as well in repeated measures regression? Thank you
Hi!
Many thanks for this tutorial. I've got a question for you. Can we use a non-normally distributed dependent variable in these analyses? My variable (anosognosia) is continuous and is kind of normal at time 0 but becomes really skewed in the following 2 waves.
Thank you
Jaime P Unfortunately, the Mixed procedure still assumes that the data is reasonably normally distributed. Although, if you have a very large data set, you can relax the assumption of normality.
@@jsparrott63 Really helpful video, thanks a lot! What would be the alternative if there is no normality then?
Dear Scott,
Excellent video, it has helped me a lot already. I just have one question: Is there an option to add an across subjects independent variable, for example, splitting the data into male and female patients and seeing if there is also a difference between gender.
I know that if the data was parametric, and if there was no missing data, I could use a 'mixed between-within subjects ANOVA' for this, but unfortunately my data is nonparametric and there is missing data, therefore pointing me towards using mixed modelling.
Any help on this would be very much appreciated.
Regards,
Tim
Yes, add that as another subject
Thank you! Very useful video!!!
Thank you. How can i control confounding variables with Generalized estimated equations analyses?
It's very blurry for me for some reason?
Thanks for the effort but audio and video quality are very poor.
this would be great, but can barely hear the audio.
Can you increase the volumn? Your voice is barely audible. Thanks.
such low resolution :(