GLM Part 6: Interaction effects: How to interpret and identify them 

Do you want to take a class with me? Visit simplistics.net to register for a class. You can either do "live" classes, where you'll learn from me directly via zoom. Or you can register for "self-guided" courses, complete with a schedule, discussion boards, quizzes, readings, etc.

Dang. this might be the only time I have come to appreciate econometrics. Thanks! You're such as great teacher!

Old video, still relevant, and yet another PhD student here eternally grateful for your ability to explain these concepts clearly! The readers of my future papers will benefit to no end because now I can actually explain my results in human words

this channel is underrated beyond belief

I don't even have finished the video and I jumped into the comments section to express my love! ahah Oh man! I'm a phD student with an Environmental science background who now is learning about statistics and I find it amazing that you can explain statistics so clearly and fun while providing such a good information. Thanks for your work man, you have a subscriber forever.

You sir are gold! Thank you for coming in at the right time!

Subscribed! Someone who truly understands their topic can explain it to someone else without jargon and this man does it well!

Super fun. I love how you break it down, delivering the content in such an exciting way that POPS! This is super helpful for learning

What an amazing teacher! Thank you for breaking it down into non-technical terms. Awesome examples.

Amazing video! Loved the way you taught the concept! Thank you!

I love your style !!! Thanks so much for your videos !!!!

You are amazing! Love your videos and how you explain complex ideas in a simple funny way

This was an amazing video, i just love how you presented this. Thank you so much!!

I am actually crying at the end of this video. I already finished my master's multilevel analysis with interactions and still coundn't explain simply WHY I needed them so bad in my model. This cleared my mind to put it in words and now I really feel ready to defend it to the panelists. ILY and thanky you, Quant Psych

omgg!! im studying this and i CANNOT understand anything until i watched your video! thank you so much!!!!!

Thanks a lot clarifying main points in such amazing manner :)

Thank you for your content, very useful and well explained! :D

Hahahaha!! What was that? Enjoyed a class after a long time! 😂😂😂 Can you make a video on mix-effect model with interaction term (within and between subject) with proper interpretation, like we write in scientific paper? And thank you again for screaming to my ear! 😂😂😂

Great video! I always come back to this one!

the best explanation of interation I ever heard!!

This is my first youtube comment ever but I can't help to express how excited I was to find this! Hands down the most fun and passionate stats videos I've seen! Everything I learned in two stats classes I took in the past year is so neatly summarized in these videos. This really helps address my doubts and concerns in conducting analysis. I can't wait to check out other videos on the channel and share this awesome resource with fellow PhD students!

awesome explanation. subscribed straight away. thank you.

You my friend are an absolutely outstanding educator. Amazing explanation.

"Now do you get it?!" "I think so 🤔" Never related so much to a video 😆

love so much your video! is clever fast, and funny, all that I ever desire to find

Thanks for the clear and funny explanation! :D

Sir, You are awesome!!! Thank you for this!

Many many thanks. Great explanation indeed

OMG!!! I FINALLY GET IT! THANK YOU! SUBSCRIBED!

the most amazing video on interactions

Thank you! Amazing video!

This is just what I needed

Thank you for the useful information

Great video! Thanks! Am I looking at the graph incorrectly, or is the introvert/extrovert graph mislabeled? The way it's labeled, introverts have greater enjoyment when there's more people.

This video is great! Thank you! You just made me not give up! Do you have a video explaining the interactions between categorical variables and how to interpret the results in an ordered logisitic regression. Or do you have especial sessions to guide this particular cases? I am studying factors associated households food insecurity, and have special interest if the household gender has an effect on the outcome and other variables.

In the comparison of Introvert & Music (~3:30), it looks like the music plots are segmented at roughly 30 people? Do people only listen to music at small parties?

Love your energy

you are brilliant!

thank you- you are AWSOME!

The introvert labelling doesn't seem right.

Omg I love this

Thanks for the useful video. Does anyone know how to plot this figure in R?

Beautiful explanation Quant Psych

thank you so much again...! 1 question...: what is the difference between a moderation model and interaction effects?

Why? Why am I only finding this channel 3rd year into my psych degree?! I've never laughed so much during a statistics lecture. I can't wait to watch more of this! Thank you, thank you, thank you for sharing your stats teaching talent!

This was super funny and super easy to understand. I would like to say Thank you but I should stop laughing first. So funny :) Thanks!

Amazing! Thank you!

Hii, thank you! I'm having a hard time: I have a regression with Y ~ YEAR * TEST(A/B). It turns out that YearB has a significant slope on the ref level (Year 1, Test A), but Test (B) doesn't. However, the interaction Year * Test is still significant (Year2, LangB, negative significant interaction), any headings on this?

Just a note on a mistake in interpretation of interaction coefficients in your paper "Eight steps...": The coefficients of a predictor (say X) in an interaction model (say Y ~ X + Z + X.Z) do not tell you the effect of that predictor X on average. It only tells you the effect of that predictor WHEN THE OTHER PREDICTOR (Z) IS ZERO. Now if you have mean-centered Z, then the coefficient of X gives the effect of X when Z is at the mean. You have given the interpretation of the interaction coefficient correctly, but screwed up the coefficients of the predictor terms.

Character In the video It's great, I like it a lot $$

Great stuff

Thank you so much !!!

Dear Professor, this was indeed a very nice video! May I ask you something? I have calculated predicted probabilities for every combination of categories of my two categorical variables (the ones that were interacted). This is the objective of my paper: Previous research has also found that health status is negatively associated with elder mistreatment (Acierno et al. 2017; Koga et al. 2019). However, social support might moderate the relationship between health status and elder mistreatment (Acierno et al. 2017). Hence, our second objective is to investigate the possible mediating effect of social support in Brazil. We hypothesize that social support mitigates the negative relationship between health status and elder mistreatment (Acierno et al. 2017). So, I want to see if support mitigates the negative relationship between health status and elder mistreatment. In this sense, my dependent variable is Elder Mistreatment (1 = Yes; 0 = No). My variable of health status is the self-rated health status, which has the following categories: Bad, Regular, and Good. And, my variable of social support is about how many family members the elder can count on, categorized as follows: None, One or Two, and Three or more. I then interacted social support with health status, and got the following predicted probabilities: * If social support is none and self-rated health is bad: Lower * Interval = 17.89; Probability = 23.68; Upper Interval = 29.47. * If social support is none and self-rated health is good: Lower Interval = 9.46; Probability = 12.49; Upper Interval = 15.51. * If social support is three or more and self-rated health is bad: Lower Interval = 8.47; Probability = 10.45; Upper Interval = 12.42. * If social support is three or more and self-rated health is good: Lower Interval = 4.83; Probability = 5.86; Upper Interval = 6.89. Based on this, could I conclude the following? There is no crossing of predicted probabilities (and their respective confidence intervals) between the levels of bad and good health, regardless of the level of social support (none vs. three or more). That is, there is at least one statistically significant negative association between health status and elder mistreatment for each of the levels of social support. Additionally, as social support increases, the negative association found occurs at lower predicted probabilities. Therefore, we have evidence that greater social support mitigates the negative association between health status and elder mistreatment, which supports our hypothesis. I'm confused if I can conclude such a thing. That is, if I am interpreting my results correctly. Many many thanks for this!

I LOVE YOU!

Dude I subscribe your channel instantly

Your amazing, 👏

Hi there. What do you mean by we are only concerned with the p-value of the interaction effect? Is this the p-value that appears in the Analysis of Deviance Table? Where does the estimate of "the slope of the interaction effect" appear? After generating a summary of my glm I get stars upon stars of p-values and Estimates (as you predicted), but all of them are for different levels of my treatments. I do not see an estimate for the slope of the interaction effect of my treatment and the second predictor variable

Hello, great video. Why not do you use VIF to know the interaction between the variables?

Great video (and channel). What happen when your third variable is not categorical but numerical?🤔

so good

Love it!, can you elaborate on why these are the two reasons we use GLM?

Perfect!!

big energy in this videos, really necessary in learning interactions! thanks

doesn't the line of best fit at 2 minutes show that enjoyment at party increases with more people attending if you are an introvert? The line slopes positively up as numbers are increasing.

My professor told me to represent my final assignment with 2 independent variables and one covariate for binary logistic regression. How can I analyse my binary logistic regression model with 2 independent variables and 1 covariate? I don’t know how can I select covariate in my data! Please help.

I'm in training for an analyst role and this video is helping me understand some key concepts. I hope you keep this up because you definitely deserve a lot more views for the number of people you are helping with this comtent.

Thanks for this nice video! But: How do we deal with several significant interactions? If my final model has 4 significant interactions, how do I interpret them? This model has then a severe multicollinearity. So, my approach is to take those 4 interactions and make 4 models for each of them for easier interpretation, but the results differ from the model, where all 4 interactions are together. I can't find anything useful on internet on this, seemingly trivial question 😩So, does anyone know how to deal with several significant interactions? Or does it make any sense? And do you by any change know some references / book, where it's clearly written and I can cite. Thanks forward!

Great explanation!!! In your face Zuur! with all respect :)

You made me laugh :)

Nice, but should you then, given AB, search for any possible variable C?

I LOVE YOU 😄😄😄

This is so help! ... But now I just need the R code for this :/

I see what you are going for with the music and the style, but (for me) could do with a little less. Most professors are drab and robotic, but this is on the other end of the spectrum. Found the music , your intensity and the constant scene cutting a bit distracting. Incredible explanation tho.

Excellent video. I'd just add that you just have to be careful about how to interpret interaction terms with both continuous variables. If you put an interaction of two continuos variables, the continuous variable that moderates the relationship is interpreted in terms of 0 and 1 (similar to a dummy). However, 0 and 1 may not make that much sense in your analysis. For instance, if you are interested in a relationship (y ~ x) that change according to age and your sample comprises people between 18-60 years old, you're measuring how your relationship change if you goes from 0 to 1 year old. This is not that useful if you're studying people between 18-60 years. Because, it can be the case that your relationship is negative for people that has zero years old, increasing until it gets positive at 10 years old. As a result, you may be misled thinking that your relationship change from negative to positive, but in your sample (18-60), it will never be negative. In other words, you can change the signal of x coefficient just changing the scale of age, because the coefficient on x always captures the effect of x when age is zero. So if you have a continuous variable interacting, I usually recommend to rescale it with mean 0 and 1 standard deviation. Then, you have a usefull interpretation: your relationship around the average value of the moderating variable is captured in the coefficient on x, and what happens with your relationship if you increase/decrease it in 1 standard deviation is captured in the interaction term.

the introvert labels have to be inverted

I love you

But but...aren't you explaining "effect modification"?

Good points but chaotic video with bad background sound

Too distracting and chaotic

your display of anxiety is giving me anxiety ...

I find the voices annoying…