Mixed Model ANOVA in SPSS with One Fixed Factor and One Random Factor

[@Luciana-wg8gb]

Good explanation. I'd just add a note or explanation on why it could be called "mixed model" and a the difference for linear mixed models, for example (afterall, this test is simply a two-way ANOVA, with one fixed and one random effect). People should be first aware of what the test is doing and the modelling rational behind it and do not mistake this for the mixed models of dependent measures (for example, a set of reapeted measures over time). The lack of standardized terminology doesn't help, but one must always bear in mind the study design and research question.

[@brednwashere]

thanks so much for posting this.. you may have saved my thesis :)

[@PB-uw7wn]

Thank-you for such a smooth explanation

[@DrGrande]

You're welcome -

[@lusummer6231]

This video is really helpful to understand mixed model .I feel confused about fixed factor and random factor, can you explain it in a simple way? Thanks.

[@AnaimK]

What if you also had multiple measurements per participant? Would 'ID' then also be a random factor?

[@pratikshasingh9199]

Hello, first of all thanks to making this video. This is very useful for my work but I have doubt about which model is good for 4 time interval and one treatment

[@Luciana-wg8gb]

Hi Pratiksha. You could use generalized estimated equations (GEE) or linear mixed models (aka mixed models, mixed effects models, multilevel models, hierarchical linear models...). Both models support a model with independent effects (treatment, in your case) and multiple measures, including different time intervals. Both models also deal very well with missing data (a very common problem in longitudinal studies). Each one is better for a particular scenario but great tools overall. :)

[@Shabbir2749]

Thanks for the video. Please explain which is the main results use to write or describe while writing ??

[@sarahdowney2136]

Are you able to provide any guidance with regards to writing this up in a report? (APA style)

[@lukefelix3687]

Hi sir! is this equivalent to nested anova? if not, how is it done in spss? been looking for vids of nested in YT but havent found one..thanks.

[@daxatak]

What is the difference between running the mixed model with this and using the "mixed model" option... In other words can you do the same mixed model statistical output with the "Mixed Models" option ?

[@davidizquierdogomez]

nice video....thanks for it. I have a question: the fix factor does not have a consistent effect in regard with city (random factor)....then, does the random effect of city remove the variability observed in the plot so we can check for the effects of the fix factor on anxiety with no troubles at all? is that why mixed effects models are used?

[@darkdistinctplaces]

Dear Dr. Grande, thank you for the video! I wanted to ask: when I run a mixed model ANCOVA, I receive non-integer numbers in df error, is that normal? Should I simply report it as a df2? I also get very strange F values and partial eta squared values compared to when I don`t use a random factor... is it normal, as well?

[@amenehshamekhi7902]

Thank you for the video. It was helpful. Do you know if there is a maximum for the levels of the random factor? I have 20 groups and when I run the same analysis in spss the Tests of Between Subjects table does not show F and sig for the factors.

[@sannijamiu6833]

Thanks so much, but what would be the model of the analysis sir

[@adithyasarma]

Hello. I had a question regarding the mixed model anova. The degrees of freedom error for main effects in your example is not a whole number. Is there any explanation as to why it is in decimals and not a positive whole number like it usually is?

[@1janez1]

Hello Tod, What if I am not interested in the interaction (not enough df)

[@rachelarney9692]

I am running a mixed-model ANOVA and do not receive an F-statistic or p-value in the output (although I do get everything else). I get an error message saying "cannot compute the appropriate error term using Satterthwaite's method". I have one dependent variable (Carbon pools), one fixed factor (4 age groups), and one random factor (12 sample sites). Any help on this would be great! Thanks!

[@yellevosteen7185]

Same problem now! similar sample size (10 sample sites) did you manage tof fix it?

[@vidb5409]

So for fixed factors, the sample size must be the same?

[@javiernavarro5760]

What about altering the syntax first?

[@unclestevescabin]

What do I do if my log transformation still yields not normalized data (my leven's test is significant)

[@usamilad]

Can you please provide me a reference for the mixed model assumption? the one that says there should be a random factor, a fixed effect, continuous response, etc I need a reference to cite why I am using this method. Best.

[@Ofel1ja]

Thank you so much!