[R] chi square test is not appropriate? but what test is

[Thomas Lumley]

On Fri, 26 Jul 2002, Petr Pikal wrote:

> Hallo all
>
> Suppose I have a summary amount of various colours used in
> different areas
>
> colour	plastic	paint
> black	15.5	173.8
> brown	6.0	523.2
> green	2.5	6.4
> red	77.1	237.4
> yellow	144.6	77.3
>
> It seems that there are some preferable colours in these two areas
> (yellow in plastic, brown an red in paint).
>
> The problem seems to me similar to chi square test but it is
> intended for testing counts or frequencies (if I understand it well).
>
> If I wanted to know whether these differences are significant,
> what type of statistical test I should use, when I have no access to
> underlaying data?
>
> Is it possible to do any statistical test on  such data?

Not without some information on the variability in the data. Even if you
were prepared to make the implausible assumption that the variance is the
same in each cell you couldn't estimate both this variance and the
interaction between color and plastic/paint.


	-thomas

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._