Elias wrote:
Some have already been answered.
> c) how formulas are changing? mean square between... (Ms) for mixed or
> not designs?
Random factors have extra variability associated with them compared
to fixed factors. This variability reflects the fact that you are
sampling a subset of items from from a population. The MSE term for
random factors has an extra component reflecting this variability.
> e) What is all about Clark and his experiment about verbs and noun?
Clark argued that some experiments sample both subjects (people) and
items (materials) from a population of possible people or items. In
this case both people and items are random factors. The obvious
example is sample words (say, nouns) from a language. If you use 30
words on a memory test and 30 people, Clark argued that to
generalize your results to other people or other words wou need to
treat both effects as random.
> f) what is the relation of quasi F with fixed and random factor?
Quasi F is the correct F ratio to use for two random effects. This
is/was hard to calculate so Clark proposed a simple solution -
calculate a minimum bound for quasi F - min F' using the F ratios
obtained from two analyses - one treating subjects as random (the
standard ANOVA) and one treating items as random and subjects fixed.
minF' can only be significant if both these F ratios are significant
But can fail to reach significance if both f ratios are
significant). I minF' tends to be conservative (but only slightly in
most cases, I think).
> j) what happen about interaction eff? i read that if we have a random
> factor we can have interaction without main effects. is right that?
Do you mean the interaction is significant and the main effects are
not? This is always possible.
> k) And generally i read everywhere random versus fixed factor issue
> but nowhere refer what is that (disadv, adv.. implications...),
> neither i found something understadable in internet or in our
> bibliothic.
Look for the commentary to Clark's orginial article, e.g.:
Clark, H. H. (1973). The language-as-fixed-effect fallacy: a
critique of language statistics in psychological research. Journal
of Verbal Learning and Verbal Behaviour, 12, 335-359.
Wike, E. L., & Church, J. D. (1976). Comments on Clark's "The
language-as-fixed-effect fallacy". Journal of Verbal Learning and
Verbal Behaviour, 15, 249-255.
Also recent articles such as:
Raaijmakers, J. G. W., Schrijnemakers, J. M. C., & Gremmen, F.
(1999). How to deal with "The language-as fixed-effect fallacy":
Common misconceptions and solutions. Journal of Memory and Language,
41, 416-426.
Thom
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at
http://jse.stat.ncsu.edu/
=================================================================
