Thanks Paul. Just want to confirm something re the random-intercept term. The point of adding this term is to soak up the between-subject variance. In other words, if we didn’t add this term, the model would still run, but the p-value would be much larger. The addition of the random-intercept term is not related to the fact that each participant is measured multiple times (in violation of the assumption of independence of observations). I’m asking because I think you said both of the above: we add the random-intercept term (1) to account for the fact that the participants will differ from each other in their responses to the predictor variable, and/or (2) because each participant is measured multiple times.
Dear, Dr. Christiansen, How would you write down this analysis in the method section of a paper? I'm struggling to write it down the correct way. Many thanks.
Great tutorial. I always wondered what is the standard in reporting lme results. I usually report them using the fixed effects F tests, but many people report the betas. Any thoughts on this?
Hi all, This tutorial is extremely helpful! Unfortunately, I am repeatedly getting the following error when trying to define my DV: "Error in eval(predvars, data, env) : object 'Cortisol' not found." Does anyone know why this would be? My experimental design is pretty much exactly like the example in the tutorial (individuals as a random factor, 3 levels of IV and a single DV). This is my first time using R software so I'm learning on the go! Thank you in advance!
Hi, thanks for the great video. I have various missing data across my dataset. Both at the response and predictor/covariate level. Is it possible to include these missing data? When I run it in R, I would assume the function would fill in my missing values/blanks, but it doesn't and therefore I get uneven groups at the different time points (outcome level). So I was wondering whether to omit the NAs or include them (and specifically can these NAs exist in the predictor level or response/outcome level or both.
Quick question Dr. Chistiansen, would this method work if there are missing variables in my data? I have read elsewhere that the ANOVA approach is not the best one if there are missing variables?
Yes, this is not an anova per se, its how you can do an anova analysis using a linear mixed effects model. A normal anova uses listwise deletion if there's missing data. This method does not.
Hi, Thanks for the lesson. It was very helpful. But I am having problem while using "emmeans". Rstudio is showing this message "Error in missDataFUN(data): argument "ex" is missing. with no default. Do you have any idea why I am getting this?
Thanks Paul. Just want to confirm something re the random-intercept term. The point of adding this term is to soak up the between-subject variance. In other words, if we didn’t add this term, the model would still run, but the p-value would be much larger. The addition of the random-intercept term is not related to the fact that each participant is measured multiple times (in violation of the assumption of independence of observations). I’m asking because I think you said both of the above: we add the random-intercept term (1) to account for the fact that the participants will differ from each other in their responses to the predictor variable, and/or (2) because each participant is measured multiple times.
This is so easy to follow - thank you!
Glad it was helpful!
Dear, Dr. Christiansen,
How would you write down this analysis in the method section of a paper?
I'm struggling to write it down the correct way.
Many thanks.
Great tutorial. I always wondered what is the standard in reporting lme results. I usually report them using the fixed effects F tests, but many people report the betas. Any thoughts on this?
Hi all,
This tutorial is extremely helpful! Unfortunately, I am repeatedly getting the following error when trying to define my DV: "Error in eval(predvars, data, env) : object 'Cortisol' not found." Does anyone know why this would be? My experimental design is pretty much exactly like the example in the tutorial (individuals as a random factor, 3 levels of IV and a single DV). This is my first time using R software so I'm learning on the go! Thank you in advance!
Hi, thanks for the great video.
I have various missing data across my dataset. Both at the response and predictor/covariate level. Is it possible to include these missing data? When I run it in R, I would assume the function would fill in my missing values/blanks, but it doesn't and therefore I get uneven groups at the different time points (outcome level). So I was wondering whether to omit the NAs or include them (and specifically can these NAs exist in the predictor level or response/outcome level or both.
Great tutorial. I wonder how I can report these results in my paper. Can you please give me some direction?
Quick question Dr. Chistiansen, would this method work if there are missing variables in my data? I have read elsewhere that the ANOVA approach is not the best one if there are missing variables?
Yes, this is not an anova per se, its how you can do an anova analysis using a linear mixed effects model.
A normal anova uses listwise deletion if there's missing data. This method does not.
Hi, Thanks for the lesson. It was very helpful. But I am having problem while using "emmeans". Rstudio is showing this message "Error in missDataFUN(data): argument "ex" is missing. with no default. Do you have any idea why I am getting this?
Do you have missing data in your data set?
@@DrPC_statistics_guides No, I haven't. It was a nominal dataset with 222 row and 4 columns.