This is utterly brilliant. Your videos have been such a great help in my dissertation. This is the answer I needed to the question "What on earth can I include as a graphical representation of my MR?" The only thing I'm struggling with now is how to write it up. But I'm going to listen to your description back and forth until I figure it out!
How would I perform what you did at 3:16 (save unstandardized predive values) with Stata??I want to grpagraphh a scatter plot with a line of fit, but I have 2 independent variables.
I have doubts about this method. Why would this method yield a positive slope for two predictors that are both negatively associated with the dependent variable? The r square and p value are correct, but the relationship is not appropriately displayed.
Hi! This is a very helpful video!! I would like to give you a question related to what you are showing. Imaging that instead of two predictors as in your video, you have three (e.g., INDEP: number of sit-ups; DEP: reaction time in a cognitive test; COVARIATES: age and IQ). And what you want to plot is the regression of number of sit-ups with reaction time adjusted by all covariates. What I do is to run a regression and ask for "Produce all partial plots". Then I have to look at the plot showing my INDEP in the X axis an all the points in the plot are adjusted by age and IQ. However, what I don't know, and that's why I am asking you for help, is where I can see the data of this plot for each participant that I have. I mean where I can obtain the values for X and Y axes. That would be great if you could help me! Thanks!
If I understand you correctly, I would regress each independent variable onto the two covariates in two separate regressions and save the residuals. Then, you'll have number of sit-ups and reaction time as two new independent variables, unaffected by the influence of the covariates. Then, conduct the multiple regression and scatter plot as I did in this video.
I would like to utilize this method of depicting the strength of multiple linear regressions, but is there a publication yet that has done this technique? Would be good to cite this as I myself am not a statistician and may not have the proficiency to justify this depiction of the relationship.
hey, thanks you mate. just a quick question, so if a use independent variable A + independent variable A on X axis and plot it againsnt dependent variable Y it would be wrong right? you cant use to variable in X axis?. the correct way is as shown... estimate and plot predict variables agaisnt dependent variable.. ? hope you can understand. greetings from brazil
Thank you for your very helpful video! I have a related query and wondering if you could help - if I wanted to show the contribution of one predictor (P1) to the overall model, after accounting for variance explained by other predictors in the model (5 predictors in total), would plotting P1 and the model predicted values in a scatter plot show this?
You would have to do it differently to this video. I think what you have in mind is a semi-partial correlation scatter plot or a partial correlation scatter plot. I discuss these two types of plots in Chapter 13 of my How2statsbook available free here: www.how2statsbook.com/p/chapters.html. In particular, I discuss a partial correlation scatter plot. I think I show how to create one at the end of this video: ruclips.net/video/W4k8Q81_470/видео.html
After watching 1,001 videos, I finally found this! THANK YOU!! Phenomenal!
Amazing video! I usually can't wrap my head around stats but this made it so easy!
This is utterly brilliant. Your videos have been such a great help in my dissertation. This is the answer I needed to the question "What on earth can I include as a graphical representation of my MR?" The only thing I'm struggling with now is how to write it up. But I'm going to listen to your description back and forth until I figure it out!
How would I perform what you did at 3:16 (save unstandardized predive values) with Stata??I want to grpagraphh a scatter plot with a line of fit, but I have 2 independent variables.
u r doing a great great JOB! I hope u gonna live long and happy.
+Paweł Iwankowski Thanks!
Superb ! Found lot of questions answer in one 5 minute 16 seconds video. Thanks a lot.
>Ishtiaque ( Bangladesh)
how do you name the x-axis after creating the unstandarized predicted value?
I have doubts about this method. Why would this method yield a positive slope for two predictors that are both negatively associated with the dependent variable? The r square and p value are correct, but the relationship is not appropriately displayed.
these are really saving me doing stats for my honors thesis! thank you!
Clever idea! I'm going to use this and see how it is received.
OMG this is awesome! I was struggling with this. Thumbs up dude! You are the best. 🤗💕
Thank you! This is so helpful! I wasn't sure how to do this.
Thank you so much, quite relevant, keep it up!
Wholesome. Thanks for this amazing video!
Hi! This is a very helpful video!! I would like to give you a question related to what you are showing. Imaging that instead of two predictors as in your video, you have three (e.g., INDEP: number of sit-ups; DEP: reaction time in a cognitive test; COVARIATES: age and IQ). And what you want to plot is the regression of number of sit-ups with reaction time adjusted by all covariates. What I do is to run a regression and ask for "Produce all partial plots". Then I have to look at the plot showing my INDEP in the X axis an all the points in the plot are adjusted by age and IQ. However, what I don't know, and that's why I am asking you for help, is where I can see the data of this plot for each participant that I have. I mean where I can obtain the values for X and Y axes.
That would be great if you could help me!
Thanks!
If I understand you correctly, I would regress each independent variable onto the two covariates in two separate regressions and save the residuals. Then, you'll have number of sit-ups and reaction time as two new independent variables, unaffected by the influence of the covariates. Then, conduct the multiple regression and scatter plot as I did in this video.
It was really amazing and useful. Thank you.
I would like to utilize this method of depicting the strength of multiple linear regressions, but is there a publication yet that has done this technique? Would be good to cite this as I myself am not a statistician and may not have the proficiency to justify this depiction of the relationship.
Thank you! Very helpful for a paper I'm working on.
Great idea, man! Thanks a lot! The method looks quite useful!
Is it possible to create risk factors adjusted scatter plots as a byproduct of multiple linear regression?
what if one of your independant variables is a categorical one, say about 5 categories of them?
Great, but how can I write the prediction equation with two independents variables, this way gives just one X??
These videos are great, thank you!
Very nice video man. Basic stuff but super interesting. Regards
hey, thanks you mate. just a quick question, so if a use independent variable A + independent variable A on X axis and plot it againsnt dependent variable Y it would be wrong right? you cant use to variable in X axis?. the correct way is as shown... estimate and plot predict variables agaisnt dependent variable.. ? hope you can understand. greetings from brazil
sweet stuff brother
Thanks - truly appreciate the tip!!
Thank you for your very helpful video! I have a related query and wondering if you could help - if I wanted to show the contribution of one predictor (P1) to the overall model, after accounting for variance explained by other predictors in the model (5 predictors in total), would plotting P1 and the model predicted values in a scatter plot show this?
You would have to do it differently to this video. I think what you have in mind is a semi-partial correlation scatter plot or a partial correlation scatter plot. I discuss these two types of plots in Chapter 13 of my How2statsbook available free here: www.how2statsbook.com/p/chapters.html. In particular, I discuss a partial correlation scatter plot. I think I show how to create one at the end of this video: ruclips.net/video/W4k8Q81_470/видео.html
@@how2stats Semi-partial correlation scatter plot - that's the one. Brilliant, thank you!
can i do this if i have two dependent variables?
Does the same method apply for a training set with three or more independent variable???
What is a "training set"?
Awesome trick! Thanks for sharing!
Hi there, I'm doing multiple regression with dummy variables due to categorical variables. Is there any way to graph this using SPSS?
what software is this?
which software you are using for plotting? Actually, I don't have any one
IBM SPSS Statistics
I have a question: Is there a way of doing this with one independent scale variable and one independent numeric variable like ethnicity or gender?
This didn't work for me. The R value of dep vs pred was lower than the r value of my multiple regression.
Which software is this???
SPSS
SPSS NOOB
love you man
thank you sir but you are doing these so fast its hard to understand quickly make it slow for other time
CAN YOU DO THIS USING EXCEL?
If you know how to calculate the predicted values with the multiple regression equation, then, yes, you could do it in Excel.
Thanks a lot!
Great !