I think you should write out your model formula and then go over it term by
term to see which term adds what to your understanding of the data. If
necessary, pick up a text about the interpretation of coefficients. For
example, you will find that Reduction is not just the slope for the
condition QF:High. It is the slope for the entire sample. This implies you
make an assumption that the slope is the same across all four conditions
that the Dialect and Prediction dummies create.
Best,
Daniel
Peter Milne wrote:
Please help me interpret the following results.
The full model (Schwa~Dialect*Prediction*Reduction) was reduced via both
update() and step().
The minimal adequate model is:
ancova-lm(Schwa~Dialect+Prediction+Reduction+Dialect:Prediction)
Schwa is response variable
Dialect is factor, two levels (QF,SF)
Prediction is factor, two levels (High,Low)
Reduction is covariate
summary(ancova) yields:
Coefficients:
Estimate Std. Error t value Pr(|t|)
(Intercept) 0.580470.09865 5.8849e-07 ***
DialectSF -0.118190.08513 -1.388 0.173315
PredictionLow -0.493090.13650 -3.612 0.000896 ***
Reduction -0.534880.15354 -3.484 0.001289 **
DialectSF:PredictionLow 0.494940.14627 3.384 0.001703 **
I know that the (Intercept) is the intercept for the first levels of both
factors (QF.High)
I know that Reduction is the slope for the first levels of both factors
(QF.High)
What i don't know is:
Is DialectSF the difference in intercepts between QF.High and SF.High?
Is PredictionLow the difference in intercepts between QF.High and QF.Low?
Is DialectSF:PredictionLow the difference in slope from Reduction?
If it's the slope, how do I find out the intercept?
With 2X2 factors and one interaction term, how many parameters am I
estimating? One slope and 4 intercepts? Or 3 intercepts and 2 slopes?
Thanks,
Peter
[[alternative HTML version deleted]]
__
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
--
View this message in context:
http://r.789695.n4.nabble.com/Help-interpreting-ANCOVA-results-tp3615352p3617024.html
Sent from the R help mailing list archive at Nabble.com.
__
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
