Re: interpretting F & t-vals from regression


Rob Campbell wrote:
<snip>.

Thankyou Bob, I appreciate your comments; I now understand what I'm doing a
lot better.

I presume you mean the meaning of the coefficients. You seem to.

But a couple of your present comments cannot be left without further
comments.


You're right that the model was over-specified but I don't think removing
intensity1000 is possible. This is a designed experiment and in the context
of this analysis, that variable is needed to explain what subjects did at
sound durations of 1000 ms.

First of all, just because you designed some level in your study
doesn't
imply what you said is correct! If a level within your design turned
out
to be useless in the light of observed data, then it's a RESULT of your
design! The removal of any redundant variable is always POSSIBLE
and appropriate.

I think you recognized the inappropriateness of your MODELING
specifications below, which is a separate matter others pointed to.


The model I was fitting wasn't intended to be predictive.

That an non--issue. Whether you meant your model to be "predictive"
or "explanatory" (which may or may not be possible), the interpretation
of the effects of whatever you have in the model remains the SAME.

To deal with this, I've run
the model again but with intensity as a continuous variable. This test
gives me one duration:intensity interaction which is not significant:
everything now ties together and is, I think, sensible.

That's a small step toward porgress. :-) But I think you're still
missing quite a bit -- in both the design and the analysis.


p.s. I know this analysis, as posted to the Group, looks a little pointless
but it's actually part of a bigger test...

As in the building of sand castles ... and long chains ... all it takes

is for the weakest link or the sand block to break for the entire
castle to crumble.

That should always be kept in mind in any statistical analysis of
a non-simple phenomenon.

-- Bob.

.

Relevant Pages

  • The Reason Why LOGOS Went Away is Obvious: Here it is
    ... There are two distinct meanings of design at play. ... This meaning of design simply says ... Here I assume you are referring to fractals as creating a design ... Adam I and Adam II - representing ...
    (talk.origins)
  • Re: The Reason Why LOGOS Went Away is Obvious: Here it is
    ... The poster who originally wrote these words about Meaning 1 and ... Intelligent Design Should Be Simple to Produce". ... natural processes produce designs such as snowflakes or fractals. ... an intelligent designer for fractal patterns made by Jackson Pollack ...
    (talk.origins)
  • Re: Making Life From Scratch, If the Creator Can Do It, It Must Be Easy
    ... There are two distinct meanings of design at play. ... Meaning 1) In one, ... at how fractals of great similarity can at multiple levels of scale. ... Adam I and Adam II - ...
    (talk.origins)
  • Re: notice and warning
    ... that script lists, edits, deletes users. ... That's just a broken design. ... POST and GET _do_ have a totally different meaning, ... keep your application safe just for the sake of using $_REQUEST. ...
    (comp.lang.php)
  • Re: Analog vs Digital
    ... > so as to propel the bird in the opposite direction. ... you have an abstract transformation - "bird wings of size S (or ... >> action, there is no design. ... > noun meaning not derived from some verb meaning of the same spelling ...
    (sci.bio.evolution)