Thanks greatly -- exactly answers my questions.
Tats

"Paige Miller" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> Okey Laboratory wrote:
>
> > I'm a biology student working to apply ANOVA models to microarray data.
I
> > wanted to brush up on my statistics before doing that, and I started
back at
> > linear modelling.  Hopefully this isn't too obvious a question that I
should
> > have picked up in my reading somewhere....  I'm currently working
through
> > Practical Regression and ANOVA using R by Faraway (it's free,
thankfully).
> > I tried to google this but didn't come up with much.
> >
> > My confusion comes from considering the coefficients of a linear-model,
say
> > in R.  What you get after fitting the model is a table of:
> > variable    estimate    standard error    t-value    Pr(>|t|)
> >
> > I understand that the t-value is the ratio of the estimate to the
standard
> > error.  After that, I'm lost though:
> > a) How do I interpret +ve vs. -ve t-values?
>
> Positive versus negative t-values ... no particular meaning other than
> the regression coefficient is either positive or negative.
>
>  > b) What do I make of the Pr(>|t|)?
>
> This is the p-value of an hypothesis test in which the null hypothesis
> is that the coefficient is zero. Basically, if you want an alpha risk of
> 0.05, if the p-value is smaller than .05, then you would reject the null
> hypothesis, and accept the alternative hypothesis, that the coefficient
> is not zero. If this doesn't make a lot of sense to you, then you need
> to brush up on hypothesis testing.
>
> --
> Paige Miller
> [EMAIL PROTECTED]
> http://www.kodak.com
>
> "It's nothing until I call it!" -- Bill Klem, NL Umpire
> "When you get the choice to sit it out or dance, I hope you dance" --
> Lee Ann Womack
>


.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to