Multiple Regression Interpretation in Excel
Вставка
- Опубліковано 25 сер 2024
- This video provides an example of interpreting multiple regression output in excel. The data set comes from Andy Field's "Discovering Statistics Using SPSS" (2009, 3rd Edition).
I've taken statistics in the past and now getting my masters and I couldn't remember hardly anything! thanks for this simple and short video to get me back into the basics!
You are an awesome teacher! I have a prof from One of the best B-schools in Europe explaining it to me like crap…got no clue of heads and tails. A 6.32 minutes of your video made sense of whole regression chapter.
Thanks a lot!
what's b schools?
Best example I've come across by far! Thank you!
Awesome video. I actually understand this now. I'm a business major NOT a mathematician so this video made so much more sense than all the equations in the textbook. Thank you!
Thanks Bella!
great explanation. Studying for a test and multiple regression makes complete sense now. Thanks!
Simple, Direct and Clear!
Great video structure. Short and to the point while packing in information. Thanks!
Great explanation. Thanks!
Thank you very much, this is much easier to understand after watching this.
I can understand your video much better compared to my ridiculously expensive and confusing statistic text book. Muchas Gracias!
:) Great to hear!
Thank you so so much for your easy-to-understand explanation!
Great explanation. Quick, to the point and every helpful.
Most helpful video I came across today. Thank you
Thanks for this. Exactly how I needed to see it explained.
Thank you for posting. Makes sense and very helpful. I can now explain my multiple regression model.
Thank You Very Much! Totally liked the explanation.
OMG! Thank you so much for this video! :)
Ugh finally! THANK YOU.
thank you for sharing this useful information with us
Ramin Ar No problem...thanks Ramin!
BEST EXPLANATION EVER!!!!!!
Really very good video...easy to understand the concepts of linear regression
Good Sir u have done a good job i have understand with help of ur 6 min lecture thanks a lot
Excellent Video!!!!
Great explanation Kavin!!!!! Explained it very well!!!!!
Great video! Very clear. Thanks!
great interpretation. Helped me a lot for my multiple regression project. Thank you!
Great video. Hard concepts for me to understand. This was helpful. You should make a series!
Thank you so much sir, it really helped a lot
GREAT example!
Excellent video and a great explanation!
great and clear explanation thank you
Thank you sir!
Glad it helped! :)
You are awesome!
This video is awesome-- incredibly helpful!! Thanks so much!
fantastic video
If they spent $1 mil in advertising and only got $288k that'd be a pretty bad deal. Right?
Of course, they might as well keep the $1 mill, since the net is $(288-1,000,000)=-712000.
Yes, very good point. Just glancing at the coefficients, it appears that for each $1 they spend on advertising, they are only getting an 8 cent return. So, yes, not a very good use of money. In this case, you would want the coefficient to be >1.00 in order to get a positive return, etc.
very good video
When doing a significance level test, which P value do you use? There are 4 P values displayed.
Thank you
Thank you sooo much! The video really helps!
very very helpful! Thank you
Thanks for this man, much appreciated
Excellent Work. Very insightful..Thanks. :)
Thanks a lot, excellent explanation, It was really clear.
Thanks so much...great explanation
Soooo helpful! Thank you
Great video!, Thank you
i have 5 IV and only 1 IV has sigificant p-value. The R square is 0.32. and in the Anova table it is 0.003 (significant)
Does these results are okay? like is the regression model ok? eventhough 4 of the IVs dont have significant P-value?
What does the standard error and the lower ans upper mean ?
Great video mate.
Amazing 😍
Very helpful! Cheers!
This is amazing!
How to do a multiple Non linear analysis in Excel ?
you are really awesome...thanks a lot.
Do you report your ANOVA? I was wondering, because it appears that the coefficients and p-value at the bottom represent more about r-squre than the ANOVA portion.
Thanks in advance...
which page is the data set in ?
sir please help me
my problem is -
Another persons explain his model -
RSEI = 0.098*NDVI + 1.019*WET - 0.025*NDSI - 0.001*LST + 0.007 (R2 =0.998)
1. The WET must increase by 0.098 if the RSEI increase by 0.1. Nevertheless, the increased WET and the decreased LST occurred at the same time .
2. Therefore, RSEI increase will be more than 0.1. If the LST decrease by 0.098 and the WET increased by 0.098 the RSEI will increase by 0.10001.
How its explain?
My model is -
RSEI = 0.9396*NDVI - 0.0074*WET - 0.22496*NDSI + 0.04665*LST + 0.0664 (R2 =1)
So,1. If the RSEI increase by 0.1 then NDVI must increase ................?
2. RSEI will increase by 0.10001, Then LST decrease by .................? and NDVI increased by ..................................?
How this model explain above mention two point?
Please answer me. and help me.
Thank you so much!!!
What is the (-26.61 variable) in row intercept and column coefficient
Thanks for the video. Is there a way of computing a confidence interval for that value of $288.43 in extra sales, please?
thank you
Any way to get Odds Ratios from the excel linear regression program?
Very Nice.. Thanx..
I didnt get why there is negative intercept. can some one explain in detail plz.
When I do the regression analysis I get data that is highlighted blue under sections like leverage, can anyone help me interpret what that means? Thanks.
Hi..What would you do if you had seasonality? Would it still be possible to make multiple regression?
Thaaaaaank you
it would be $288,430 on sales a week, month, or year???? thks
is it not 95% confidence level?
TheWoundedDoctor Thanks a lot, 'wounded' doctor :)
What's the lower and upper 95%??
Hey Kevin,
I'm just getting started with multiple regression and your video has been a great help! However, I am wondering how to interpret the high standard error? As far as i know this indicates that there is a high correlation between the independent variables, right? Could you elaborate on that? Is it a problem?
Isabella Nocon Hi Isabella, thanks for this. Standard error is basically the standard deviation of the sampling distribution. It does not relate to the correlation between independent variables. Basically, a high standard error (standard deviation) makes the test statistic lower, which can make the p-value higher. One of the ways to lower the standard error is to increase your sample size (this makes sense if you look at the formula for the standard error).
As far as exploring correlation between independent variables (i.e., 'multi-collinearity')--one easy test is to do a cross-tabulations table, and simply explore the correlation coefficients between the variables.
Kevin Brown Hi Kevin, thanks a lot for your response! I followed your advise and my correlation coefficients were at 0,8 and 0,86 so I guess there is indeed a strong multicollinearity between my independent variables. However I've read that one of the basic assumptions about multiple linear regression is that no multicollinearity is permitted. Does that make my model now automatically insignificant?
Isabella Nocon Hi Isabella, while those values are high (.70 seems to be a common cutoff) that does not automatically make your model insignificant. As I understand it, you have to recognize what is wrong with multicollinearity. Specifically, it can make your standard error artificially inflated (which can ultimately artificially inflate your p-value). In other words, it makes the results worse than they actually are. However, if you still have p-values that are low enough to claim statistical significance, then it shouldn't be a problem. Now, some statisticians may disagree and simply say that those levels are too high--but again, if your p-values are low enough, then it is not as great of a concern (NOTE: Conversely, this is why heteroskedasticity is such a problem: it does the opposite. It artificially lowers the standard error and makes the model look better than it actually is).
Kevin Brown Ah ok, now I get it! Thank you very much for your patience and helping me out here :)
Isabella Nocon No problem!
hi thanks for sharing this video but i couldn't understand why it is statistically significant at 99% not %95? it seems like you used 95% confidence level?
ayla serin It is statistically significant at the 99% level and the 95% level (the independent variables, that is). That is because the p-value (which is the probability of committing a type-I error--or rejecting the null hypothesis when it is in fact true) is less than .01 (and .05).
^_^ awesome explanation !!! ... I salute you
you have not explained about standard error
Please show your work next time, that would be very helpful!!!
"You're too ugly"... Stats can be fun
I think he's sick