Coefficient of determination (r-squared)

In a linear regression, you often see the R-squared quoted. To explain the R-squared (coefficient of determination), I compare it to the standard error of estimate (a measure of the line's accuracy) and the correlation (the square root of the coefficient of determination). All three, loosely speaking, are measures of the line's fit to the data

Category:

Education

Tags:

• Finance
• Quant
• Excel

[nozy03]

I have r^2= 0.37. Would you say 37% is good?? I'm not sure what to write.

[bionicturtledotcom]

@nozy03 Ideally, you would test the significance of the R^2 (akin to t test of slope coefficient). While some have a high standard, i think high R^2 are can be unrealistic (most variables are informed by many variables). In my humble opinion, 0.37 conditional on significance, is quite meaningful

[whkrause]

Um, independent and dependent variables only exist in an experiment. Are you saying you ran an experiment where you manipulated Yahoo stock prices to see its effect on Google stock prices?

[bionicturtledotcom]

@whkrause No, no "effect" implied. The prices are observed as samples drawn from a (theoretical and unobserved) population. Variables don't require, to my knowledge, an experiment: they merely reference the formula which can be used on any dataset

[adeeleo]

On what basis Google is chosen as dependent variable on y-axis and Yahoo on x-axis as independent variable?

[openuniverse2003]

Excuse me, at time mark 8:00 to 8:06, GOOG s/b the DEpendent variable, no?

[nozy03]

@bionicturtledotcom Wow thank you for your quick reply. I am an undergrad 1st year psychology student and haven't yet been told to test the significance of R^2 (although, now I have seen your video I may just shove it in there!). Was just nice to find a second opinion on 37%.

Enjoying your videos, just subscribed. Thanks very much!

[thehiral84]

thank you.



[macdean]

thanks. What kills me is my Wooldrige book (p 41) calls SSE the explained sum of squares and SSR the residual sum of squares. Your tutorial calls is the Sum Squared Regression (at 5:32) which is slightly different than Sum Square of residuals.It makes it all the harder to understand when statisticians don't agree on the convention.

[ongolos]

yahoo suxs , google should be x, not y.

[unit45x]

Thank you so much! Very helpful!!!

[tufstuf22]

THANK YOU FOR POSTING THIS! my AP stats book does a really horrible job of explaining this so thank you!