Modern repeated measures analysis using mixed models in SPSS (1)

Repeated measures analyse an introduction to the Mixed models (random effects) option in SPSS. Demonstrates different Covariance matrix types & how to use the Likelihood ratio test to evaluate different models.

Robin Beaumont
Full notes, MCQ's etc at:

[dtolin]

Thank you for posting this. Do you happen to have an example of a mixed-factor analysis (e.g., comparing two treatments over time)?

[augustus]

Hello, is mixed model a parametric or a non-parametric analysis? In other words can I use it when my data is not normally distributed?

[umaralhajiabubakar]

These videos are great. They are really, really help-full. Thanks.

Umar Alhaji

[TommasoMazza]

Hi, how do you compare repeated measures with missing data?

[dirkschomburg]

at 8:46 the table with Covariance matrixes. Is there an error? with "Diagonal" there should be "No correlation (i.e. independent)" in the last column. If I compare that with the IBM SPSS documentation.

[bhattacc]

sorry it is me again, i have a data with one withing group IV with 3 levels  and one between group iv (which is a continuous variable) and i need to do repeated measures regression, this tutorial covers what do with the within group IV, how should I handle the between group iv in the mixed model? is there a tutorial for that in your link? Thank you for your help again.

[nikoschristodoulakis]

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?

[ruslanvolkonsky]

This is really helpful.
Can you show a 3-way mixed effects model please?
For example:
IV1 = Gender (btw)
IV2 = Time (before or after learning)
IV3 = Type of learning (with say, 5 levels)
DV = recall

[saho]

Thank you. How can i control confounding variables with Generalized estimated equations analyses?

[qwertY]

Hi Robin, Thank you for this clear video. I have a question about the modeling of change scores. I want to compare change in the dependent variable between two groups and they differ in baseline scores. Do I have to control for this difference, or does MLL account for it automatically. I thought that it accounts for this change as it models change withing groups and compares this change between groups but now I'm confused. Can you help me? Anyway, many thanks.

[thomasp]

This is by far the best explanation of mixed models I've seen online. I'm regressing a repeated measures continuous IV on a continuous DV and would like to make predictions, thus I need the regression equation and (ideally) the R2. Can I use the mixed model to get this?

[Makroragerovia]

Hi,

I would like to ask about the model comparison at the end of the video. What is the interpretation of 1-pchisq (= 96)? Is it that 96 of 100 times the models would be equal in fitting or that 1 of 96 is different? So if the first is true, than the 1-pchisq should be higher thand 0.95 to conclude equality, right? In the second case (1 of 96 different) it schould be huigher than 20 (giving 95%). Thank you! M.

[bhattacc]

thank you so much I have the same question as Sarah garner

[zubinbhavsar]

can u show how to do 2x2x8 mixed design anova please.

[RobinBeaumont]

there is an example just like this as part of the SPSS mixed models tutorials.

[wifflek]

Hello there, thank you for this really useful video. I am trying to run this analysis myself, however I have the equivalent of 3 time points (instead of your 2). Would you be able to explain how to do this? Thank you.

[bachadega]

Would you do an example where apply mixed model to marketing activity...for instance, to find the coefficient for a total sales of a company which is dependent on some stores and these stores sell five of the company's products..highly appreciated!!

[RobinBeaumont]

Hi chandrima 
I would suggest that you have a look at the book Linear Mixed models by West, Welch and Gatecki - 2nd ed. (2014) this has a example of what you want in it - I assume you are using R

[brunomazuquin]

Hi Robin,
This is a great video, thanks for sharing your knowledge with us.
What are the references you have used to explain the choice for the covariance matries for repeated measures?

[pal]

Very nice series. I wonder if you'd be interested in doing a similar set of videos for binary or count outcomes? Thanks again.