Linear mixed effects models

Рет қаралды 218,931

Күн бұрын

When to choose mixed-effects models, how to determine fixed effects vs. random effects, and nested vs. crossed sampling designs. Requirements and assumptions of mixed-effects models, and how to evaluate them. How mixed-effects models can improve parameter estimation with partial pooling/shrinkage.

Пікірлер: 47

@fiore1394 8 ай бұрын

Oh my goodness, thankyou for making a video that actually explains statistical content clearly! If I had a dollar for every video with a title like, "such and such analysis method, CLEARLY EXPLAINED!" then goes on to dive into the most complex content imaginable without proper explanation I'd be a very rich man. Sorry about this vent, I'm just very appreciative. Keep up the good work.

@animamundii 3 жыл бұрын

By far the best explanation on LMM. Thanks

@nicolasmellalopez3684 2 жыл бұрын

Great explanation man, I really appreciate the effort! Although there is a lot of information available and also a lot of sources where to find them, it takes a lot of effort to explain these kind of models graphically. I've read about these models from 2 or 3 different sources in order to get a general picture, but this one is a nice and clear explanation, besides been shown as figures

@zheyuanpei5543 Жыл бұрын

I am new to this model and I have to say that this video is really helpful! Thanks!

@muffinsniper 3 жыл бұрын

Studying psychology and this was super helpful!! Thanks

@kersting1712 Жыл бұрын

this is the best explanation I saw so far! thank you so much!

@Maicolacola 3 жыл бұрын

This was incredibly helpful. Thank you!

@gavinaustin4474 4 жыл бұрын

Thanks Matthew. Very good explanation.

@gisellechuwen2236 2 жыл бұрын

thank you so much! this is so helpful and you are great at explaining.

@srishtigureja6534 2 жыл бұрын

This was really helpful. Saved my day!!

@mallorythomas725 Жыл бұрын

Really good explanation! Helping me write my first manuscript :)

@amalnasir9940 2 жыл бұрын

Thank you sir! Even with this simple explanation the topic is still complicated. I wished the examples were simpler.

@multitaskprueba1 11 ай бұрын

Fantastic video! Thank you so much! You are the best!

@sean_gruber 3 жыл бұрын

Great video, thanks!! Just enough information to get me started without going into full-blown detail.

@ernstpaulswens 3 жыл бұрын

thanks! very clear visualisations

@stevenash2869 Жыл бұрын

The bext explanation I've found, thank you!

@rohanshinkre 3 жыл бұрын

Sir thank u so much 😊 best explanation period

@lintonfreund Ай бұрын

this video is incredible, thank you so much!

@user-mh7px2uy1k 7 ай бұрын

Excellent work

@marco.miglionico Жыл бұрын

Great video

@y37chung 4 жыл бұрын

Great video. I have a question, what would make more sense to be used for accounting inherent agricultural field variability (having spatially separated block on a larger field)? A fixed or random effect?

@francycharuto 3 жыл бұрын

Random

@bezaeshetu5454 3 жыл бұрын

nice explanation. Thank you for posting. Can you share some materials on GLMM please? Thank you so much it really helps.

@jcmt8178 3 жыл бұрын

Thanks Matthew. In a longitudinal design, let's say 5 Time Points, 20 subjects what would be the optimal way to set up the random effects? I feel like whenever I include the intercept or any interaction with tie TIME POINT factor it explains almost all variance in the dependent variable (as it changes from time point to time point, but I want to study the effects of the independent variables changing over time on the dependent variable). Should I just ignore the TIME POINT (or "visit "1, 2, 3 4, 5) factor, as it's implicitly related to the values both in the dependent and independent variables? And just include the "SUBJECT" as a repeated measures account?

@maxgav13 4 жыл бұрын

Thanks! Great explanation and summary. I wanted to ask if there's a source (paper, book, books) you could point to for this topic? Thanks again

@madisonlong3182 3 жыл бұрын

Just to update anyone else who comes looking for a citation, the manuscript Naseem linked was recently published in Advances in Methods and Practices in Psychological Science! journals.sagepub.com/doi/10.1177/2515245920960351

@ericdoe1129 3 жыл бұрын

Great

@milanfilipovic3648 2 жыл бұрын

how is partial pooling or shrinkage model different then running a fixed effect model on that subset of observations?

@Jillllllllll 13 күн бұрын

SUPER nice!! one question, i have a LMM with df what do they mean?

@driouechehocine535 2 жыл бұрын

Hello and thank you for the video I would like to use GLMM multinomial logistic regression mixed model for repeated data with R software, response ~ trt + period + seqTrt + (1|id) do you know a package or a function for this model thank you in advance

@will74lsn 2 ай бұрын

can I find somewhere examples of random coefficient models where the variable of the random coefficient is not continuous but categorical? ideally written with STATA or SPSS?

@JinaneJouni 2 жыл бұрын

Can someone help me to do the plot where we visualize the lines with different intercept and slope? I'm using Rstudio

@chacmool2581 Жыл бұрын

Country X has 30 states with repeated observation measures of X across 15 years for each state. Is Mixed Effects appropriate to model Y from X with states as random effects?

@divyaagrawal6740 Жыл бұрын

if I have the more than 3 datasets with different x and y axis then how statistically it can be compared??

@ufoisback5088 3 жыл бұрын

The R code for all this stuff would be great

@ghadaelkhawaga7081 2 жыл бұрын

How can I write a comment on mixed linear model plz

@a.s.3874 3 ай бұрын

Are LMM and LMEM the same thing?

@ViriatoII 3 жыл бұрын

Awsome explanation. But wait, I just can't get p-values? How do I know which fixed effects are relevant?

@MatthewEClapham 3 жыл бұрын

It may depend on the program you're using, but the authors of the lmer function (lme4 package in R) chose not to give p values. However, there are standard errors for each coefficient and you can get the 95% confidence interval on each fixed effect by running the confint() function on the model output.