# Thread: increasing the r squared

1. ## Re: increasing the r squared

thanks im getting this now

2. ## Re: increasing the r squared

One thing to remember, and it is easy to forget, is that you can get everything right and still have poor R squared values. If the pheneomenon is really complex no small set of variables may in fact explain much.

3. ## Re: increasing the r squared

whats considered poor ideally i would like 1, but i know that aint happening also is the durban watson test worth looking at in terms of being over 0.8 for auto correlation

4. ## Re: increasing the r squared

Do you have time series data (normally the only reason you check Durbin-Watson).

5. ## Re: increasing the r squared

whats considered poor ideally i would like 1
I don't know of anyone who has ever achieved this with empirical datasets. As to what is considered poor, that depends on your system. In physiology, where patterns of covariation are typically governed by tight constraints (e.g. receptor binding affinities, cardiac output, etc.), "good" r-squared values are likely to be high. For ecological data (e.g. influence of rain on mating probability of frogs), "good" r-squared values might be very low (0.25). This is why people typically don't examine r-squared during the model simplification process. We just look for the best model for the data we have, assuming that the data we collected can adequately address the question on hand. Even adjusted r-squared (which accounts somewhat for model complexity) has been shown to perform worse than AIC for selecting the simplest and best explanatory models, especially for data with nonlinearities.

You're smart to look for autocorrelation in your data. I would be more interested to know if you have repeated-measures data (more than one data row or measurement per individual).

Good luck. Sounds like a cool data set!

6. ## Re: increasing the r squared

jpkelly makes a really good point. What is a good value depends entirely on what you are measuring. Certain phenomenon are very complex (meaning many factors contribute to them) and thus will have much lower r squared values than other phenomenon. The best idea is to look at journals in the area, or trade publications, and see what types of r squared values they have for similar studies.

7. ## Re: increasing the r sqaured

Originally Posted by noetsi
I always thought that if you transformed the Y value to a log you had to do the same to all predictor variables.
I don't think this is true. But then again, I guess it depends. In what kind of situation would you transform all the predictor variables?

8. ## Re: increasing the r squared

I thought, although it appears that I was wrong after reading more, that since you logged the left hand of the equation (the dependent variable) you had to log the right hand of the equation as well (the predictor variables). Much as if you multiply one side of an equation by a constant you have to multiply the other side.

But again I find my long assumed views are wrong in this regard....

9. ## Re: increasing the r squared

thanks all for the help

10. ## Re: increasing the r squared

You are welcome.

Incidently what is really important is getting the model right not getting a high r squared value

11. ## Re: increasing the r squared

Incidently what is really important is getting the model right not getting a high r squared value
Yes, noetsi is right! Very important!