Man, you've saved my assignment- not only that but you've helped me *actually* understand what I'm doing, rather than just following along with unexplained instructions like in class practicals!
i watched your video twice professor, and it really helped me writing and concluding my final desertion. Really worth watching, easily understandable, my respect for you!
What a nice surprise! I'm working on a statistics project right now and typed how to check assumptions into Google, clicked on the first result. Didn't think it would be one of my favorite UA-camrs!
I have returned to this same video about a hundred times, it is awesome! I am wondering if you would be able to make a video showing how to do a ridge regression in SPSS. there are no videos showing how to use this new function!
I found this video easier to follow. Thank you Dr. Grande for the detailed information. It was very useful the explanation on how to conduct the assumptions and especially the values, what to look for and the explanation of it.
Awesome video that explains the multiple regression in a clear step by step manner. My only remaining concerns is what if my data deviated from the set assumptions (e.g. mulicolinearilty; normality). The data used seem perfect. What if it wasn't?
It was interesting when you pointed out the durability of treatments and how symptoms decrease when treatment continues but not so much when treatment stops. This is a good research topic.
For people not familiar with linear algebra independent variable is what statisticians use typically on the x axis and y axis is outcome of straight line y=B*x aka dependent variable.
This is much easier tan what I was taught to test assumptions and see if you are on the right track! Thanks! A for my real analysis, lol for a moment I thought I wouldn't have to dummy code, lol silly me ;-)
Thank you for explaining a difficult issue in such a simple way Dr. Grande. It was very helpful for me. I want to ask that if one of these assumptions is not met (i.e. multicollinearity of the variables) is there any way to correct this condition? Or should I renounce from Lineer Regression and do another analysis?
Sir I have three variables. 1. Academic burnout 2. MENTAL health 3. Self efficiency How hypothesis and objective is that how Academic burnout impact our mental health and Self efficiency so in this case what our IV and DV and which one should we use linear or multiple regression? Please sir respond 🙏
There are a lot of details in this video. I think it would be very good to follow along if/when I need to perform a multiple linear regression. For now it is a bit much to really follow.
Dear Dr.Todd, your video is the very helpful to me, what consequences do I have if my std.residual min is -3.790, thank you so much in advance if you could help me 🙏🏼🙏🏼🙏🏼
Thanks for this informative video! Can I just check, youve mentioned for the normality tests if the p value falls below .005 then it violates normality assumption, however so far at uni ive been told if they fall below 0.05 (not .005) then it would be considered a violation. Which value should we use as a guide?
Thank you for explation. I am very happy that there are videos about spss program and analysis such is multiple regression that help to learn and understand about this important part of analysing results. I have just one question about part correlation, can we interpret it as unique contribution of predictor as procentage when we multiple it with 100 or we first must square it and then multiple it with 100? Thanx for effort, all good, God bless you
Mr. Grande, you mentioned the correlation values as one of the assumptions. But I could not find any reference to the correlation value between dependent and independent variables greater than .30. I would be very glad if you could help with this.
Thank you. What if we find that the dependent variable is not normally distributed. How can we continue? Also can you please refer me to anothet linear regression video of yours, if any, where one of the independent variables is nominal?
Thanks Dr.Todd. We read in statistical books that regression residuals have normal distribution. But you check normal distribution of dependent variable. which one is true?
I heard another scholar used the partial correlation as the covariance each predictor explains uniquely when the other predictors are held constant. How that is different from using the part to explain the unique contribution of each predictor?
Thank you so much Dr. Todd for your videos. They are very helpful. Please, can a categorical variable eg gender, socioeconomic class be used as one of the predictor variables in multiple linear regression?
Hey Dr. Grande! Can you give me a list of references for the range of acceptable values to judge multicollinearity and correlations that you mentioned here? Thanks!
Thank you for your video! Very clear and helpful. I have some questions. I have an independent variable that has a correlation below .3 with the dependent variable, how does that affect interpretation? It's .265. Also, my data is not normally distributed, but I'm using a sample that is reflective of a population and asking about rare behaviours, thereby diversions from normality are likely to represent difference in the population. Would that mean that I am okay to go on and interpret my results despite lack of normality?
I want to thank your for all your videos, I have a query on this one. I have not met the assumption of normality. There is one record that seems to be throwing it out. Should I remove the record?
i am wondering, when checking the residuals table, the standard residual cannot be lower than -3. what happens if it is? which assumption is violated and how can i solve it ?
Todd Grande In the video you're talking about checking if dependent variable is normally distributed. As far as I understand, linear regression doesn't care about dependent variable distribution, it doesn't assume its normality. What can be assumed is normality of errors distribution, and even that is an optional assumption, made just for convenience.
Dr. Todd Grande, thank you a lot, this was very helpful. I have a question, can I make a multiple regression analysis with independent variables, which are different on scales, for example, one questionnaire subscales (the subscales mean are about 2-3) and mathematics examination, which mean is 70 percent?
Thanks so much for this video. When carrying out my regression, my scatterplot has variables above and below 3 and -3, respectively - the standard residual range is -4.450 to 6.184. My dataset is about n=5500 but there are multiple values outside the -3 to 3 range. When trying to find explanations for this, I came across a video that suggested heteroscedasticity but I'm not sure if it was applicable to my case or not. Are there any tips or help you could provide about this situation? Thank you again for the video
Hi Dr. Grande. I have a question. Knowing that when the independent variables are at individual level of analysis and the dependent variable is at group level of analysis the levels of analysis have to match before we perform a multiple regression, what is better: (1) To divide the group score for the dependent variable by the number of individual cases in a group and give each case an individual score for that dependent variable, or (2) To add up the independent variable scores of all members in a group and use that score to conduct the analysis on the group level of analysis. However, when we do (2), the sample size decreases, which creates problems. What do you think?
Dr. Grande, thank you very much for the video. One thing I wanted to ask is whether we are interested in the normality of residuals of the model, and not the raw scores?
Thank you Dr for your valuable video. My question is if my survey has 3 independent variables and one dependent variable, where each of them is assessed by about four questions in the survey, I.e. the survey has 16 questions. So, how can I do the multiple regression analysis in this case. Could you please assist me. Your prompt response is highly appreciated. Mohamed
But the problems i have created in the last couple hours are detrimental for my personal life. I have created the decision whether I choose to be gay or hetero in SPSS, and the multivariate regression analysis shows the predicition of being gay are way higher then hetero, what should I do? Please someone help me
Dear Dr. Grande, would you please update (e.g., Videos of 2021, 2020, 2019, etc.) your lectures (e.g., Stepwise multi regression analysis, etc.)? I wish to add them to my final dissertation if I have your permission.
dear sir, thanks for the video. i have a doubt. in the coefficients table, the constant is our B0 and the values written below it (instandardized) are our B1right? This B1 value is then written in Z-score terms in the Standarized beta column. right???
Man, you've saved my assignment- not only that but you've helped me *actually* understand what I'm doing, rather than just following along with unexplained instructions like in class practicals!
i watched your video twice professor, and it really helped me writing and concluding my final desertion. Really worth watching, easily understandable, my respect for you!
you've saved the day for me, doing my dissertation proposal and much more confident about everything now
What a nice surprise! I'm working on a statistics project right now and typed how to check assumptions into Google, clicked on the first result. Didn't think it would be one of my favorite UA-camrs!
This was amazing, thank you so much for not only running the test but also clearly interpreting the results.
Thank you, Dr. Todd Grande. Your lecture really helps me do my dissertation.
SOOOOO helpful. thank you for this amazing free learning resource!
Thank you so much. Your old video still serving so many students around the world. Keep up the good work.
> Ishtiaque (Bangladesh)
Stay blessed Dr. Todd Grande for your several lectures on SPSS. I am happy with your efforts Sir.
This video has been so helpful in many ways. Thank you so much Dr Todd.
Thank you, the instructions in the video were clear and helpful. Hope that you will have a good day.
Dr. Todd Grande is always the best!!!!!!
Stay blessed Dr. Todd Grande for your several lectures on SPSS. I am happy with efforts Sir.
I have returned to this same video about a hundred times, it is awesome! I am wondering if you would be able to make a video showing how to do a ridge regression in SPSS. there are no videos showing how to use this new function!
That is really great to find u here dear Prof...u have enlightened my way towards my hillarious R.P, thk u so much.
Greetings from Kurdistan-Iraq
I found this video easier to follow. Thank you Dr. Grande for the detailed information. It was very useful the explanation on how to conduct the assumptions and especially the values, what to look for and the explanation of it.
Thank you very much for your great online lectures.
Very useful, thanks for creating this video.
Thank you very much for this detailed explanation!
Awesome video that explains the multiple regression in a clear step by step manner. My only remaining concerns is what if my data deviated from the set assumptions (e.g. mulicolinearilty; normality). The data used seem perfect. What if it wasn't?
it is very interesting explanation how to test regression assumption test. stay blessed
thanks, you're the real MVP
I m so thankfull to you. May God bless you Sir.
thanks dr grande i hope to learn more from you
It was interesting when you pointed out the durability of treatments and how symptoms decrease when treatment continues but not so much when treatment stops. This is a good research topic.
Thank you very much! I got a refresher on regression!
For people not familiar with linear algebra independent variable is what statisticians use typically on the x axis and y axis is outcome of straight line y=B*x aka dependent variable.
This is so helpful. Thank you so much 🙏
Excellent explanation.
What do we do if the normality tests are significant?
Yes, excellent video, there are many assumptions but I got it, thanks.
Helpful review of Coefficients. Thank you.
Thank you for uploading it!
This is much easier tan what I was taught to test assumptions and see if you are on the right track! Thanks! A for my real analysis, lol for a moment I thought I wouldn't have to dummy code, lol silly me ;-)
Excellent
Thank you for explaining a difficult issue in such a simple way Dr. Grande. It was very helpful for me. I want to ask that if one of these assumptions is not met (i.e. multicollinearity of the variables) is there any way to correct this condition? Or should I renounce from Lineer Regression and do another analysis?
Useful video. Thanks,
What do I gonna do if I don't have a normal distributed in a dependent variable !!
Sir I have three variables.
1. Academic burnout
2. MENTAL health
3. Self efficiency
How hypothesis and objective is that how Academic burnout impact our mental health and Self efficiency so in this case what our IV and DV and which one should we use linear or multiple regression? Please sir respond 🙏
Hi doc,.. what happens and what should we do (and how) if the dependent variable data is not normal?
I'm also curious about this.
how do we run this if the DV is not normally distributed?
There are a lot of details in this video. I think it would be very good to follow along if/when I need to perform a multiple linear regression. For now it is a bit much to really follow.
Dear Dr.Todd, your video is the very helpful to me, what consequences do I have if my std.residual min is -3.790, thank you so much in advance if you could help me 🙏🏼🙏🏼🙏🏼
thank you sir really helped a lot
Thanks for this informative video! Can I just check, youve mentioned for the normality tests if the p value falls below .005 then it violates normality assumption, however so far at uni ive been told if they fall below 0.05 (not .005) then it would be considered a violation. Which value should we use as a guide?
Thank you for your knowledge.
Thank you for explation. I am very happy that there are videos about spss program and analysis such is multiple regression that help to learn and understand about this important part of analysing results. I have just one question about part correlation, can we interpret it as unique contribution of predictor as procentage when we multiple it with 100 or we first must square it and then multiple it with 100? Thanx for effort, all good, God bless you
Thanks for this helpful video
Thank you professor...
Simple and clear
Very Informative, I learn a lot
I'm glad you found the video useful. Thanks for watching.
Thank you for a great video!!
You're welcome -
very good explaination. thank you.
I'm glad you found the video useful. Thanks for watching.
Mr. Grande, you mentioned the correlation values as one of the assumptions. But I could not find any reference to the correlation value between dependent and independent variables greater than .30. I would be very glad if you could help with this.
I have a Kolmogorov-Smirnov significance of .200 as well but a Shapiro-Wilk significance of .003
Thank you. What if we find that the dependent variable is not normally distributed. How can we continue? Also can you please refer me to anothet linear regression video of yours, if any, where one of the independent variables is nominal?
Thank you
thank you so much that was very clear
You're welcome!
Thankyou very much sir
thanks Dr Grande. Could you please make a video on how categorical variables can be used in the model?
Thanks Dr.Todd. We read in statistical books that regression residuals have normal distribution. But you check normal distribution of dependent variable. which one is true?
I heard another scholar used the partial correlation as the covariance each predictor explains uniquely when the other predictors are held constant. How that is different from using the part to explain the unique contribution of each predictor?
i need a reference for the idea that a corr. smaller that 0.7 is enough to verify the lack of multicollinearity assumption.
Thank you for sharing :)
Hello Dr. Todd,
Thanks for this lovely video. I would like to do this analysis step by step with you. Where can I find the data set?
Thanks.
Thank you so much Dr. Todd for your videos. They are very helpful. Please, can a categorical variable eg gender, socioeconomic class be used as one of the predictor variables in multiple linear regression?
1 years later, sorry. Yes, it can be used.
Thanks
Sir, in case of multiple linear regression, normal distribution has to be checked only for dependent variable or entire variables
Hey Dr. Grande! Can you give me a list of references for the range of acceptable values to judge multicollinearity and correlations that you mentioned here? Thanks!
Dr Grande must be awarded for lowest volume videos on youtube :D jk
Dr. Todd, thanks for the nice video, can I ask one question, how can we obtain 95%CI of β ?
Thank you for your video! Very clear and helpful. I have some questions. I have an independent variable that has a correlation below .3 with the dependent variable, how does that affect interpretation? It's .265. Also, my data is not normally distributed, but I'm using a sample that is reflective of a population and asking about rare behaviours, thereby diversions from normality are likely to represent difference in the population. Would that mean that I am okay to go on and interpret my results despite lack of normality?
Hi, what are your sources for the assumptions? Thanks in advance! :)
I want to thank your for all your videos, I have a query on this one. I have not met the assumption of normality. There is one record that seems to be throwing it out. Should I remove the record?
Welke assumptie wordt geschonden als de correlatie tussen de onafhankelijke en afhankelijke variabele lager is als 0.3?
i am wondering, when checking the residuals table, the standard residual cannot be lower than -3. what happens if it is? which assumption is violated and how can i solve it ?
Thank you very much !
You're welcome -
Todd Grande
In the video you're talking about checking if dependent variable is normally distributed. As far as I understand, linear regression doesn't care about dependent variable distribution, it doesn't assume its normality. What can be assumed is normality of errors distribution, and even that is an optional assumption, made just for convenience.
Whats the best way to identify the skewed independent variable if you have loads of variables
Outcome variable need not have a normal distribution rather error terms should be normally distributed.
How to conduct linear regression analysis when the variables are not normally distributed?
Mumtaz ! = Excellent !
What is the difference between unstandardized and standardized B coefficient ?
Dr. Todd Grande, thank you a lot, this was very helpful. I have a question, can I make a multiple regression analysis with independent variables, which are different on scales, for example, one questionnaire subscales (the subscales mean are about 2-3) and mathematics examination, which mean is 70 percent?
My mean is not around 50... Should it be? I didn't understand this part
Thanks so much for this video. When carrying out my regression, my scatterplot has variables above and below 3 and -3, respectively - the standard residual range is -4.450 to 6.184. My dataset is about n=5500 but there are multiple values outside the -3 to 3 range. When trying to find explanations for this, I came across a video that suggested heteroscedasticity but I'm not sure if it was applicable to my case or not. Are there any tips or help you could provide about this situation? Thank you again for the video
Hi Dr. Grande. I have a question. Knowing that when the independent variables are at individual level of analysis and the dependent variable is at group level of analysis the levels of analysis have to match before we perform a multiple regression, what is better: (1) To divide the group score for the dependent variable by the number of individual cases in a group and give each case an individual score for that dependent variable, or (2) To add up the independent variable scores of all members in a group and use that score to conduct the analysis on the group level of analysis. However, when we do (2), the sample size decreases, which creates problems. What do you think?
What should I do if assumptions are violated, i.e. normality or linearity?
I was under the impression only the residuals had to be normally distibuted, not necessarily the dependent variable. Is that right?
Dr. Grande, thank you very much for the video. One thing I wanted to ask is whether we are interested in the normality of residuals of the model, and not the raw scores?
I am also confused about this. I was under the impression that it was just the residuals, not the raw data of the variables themselves?
Thank you Dr for your valuable video.
My question is if my survey has 3 independent variables and one dependent variable, where each of them is assessed by about four questions in the survey, I.e. the survey has 16 questions. So, how can I do the multiple regression analysis in this case.
Could you please assist me.
Your prompt response is highly appreciated.
Mohamed
But the problems i have created in the last couple hours are detrimental for my personal life. I have created the decision whether I choose to be gay or hetero in SPSS, and the multivariate regression analysis shows the predicition of being gay are way higher then hetero, what should I do? Please someone help me
Hello Dr. Grande! This is awesome! Do you have a sample on how to write all of this APA?
Is it important to test normality for independent variables?
Dear Dr. Grande, would you please update (e.g., Videos of 2021, 2020, 2019, etc.) your lectures (e.g., Stepwise multi regression analysis, etc.)? I wish to add them to my final dissertation if I have your permission.
dear sir, thanks for the video. i have a doubt. in the coefficients table, the constant is our B0 and the values written below it (instandardized) are our B1right? This B1 value is then written in Z-score terms in the Standarized beta column. right???
Do we not check for the Normality of the Residuals instead of our dependent variable itself? Does the DV need to be normally distributed as well?
OH MY GOD ARE YOU ARIANA GRANDES FATHER ??? THAT IS AMAZING
greetings Doc. do we have to ck on normality if i am using a census method considering the sample size is few in no. like #111?
How to handle the data if correlation between IV and DV is less than .3. I can't remove my IVs.
If the data was Shapiro Wilk test significant how the further regression will be carried out?
thank you.
Dr. Grande, can you suggest a reference(s) for the bounds on the parameters that you suggest in the video?