[R-lang] Re: Test for binomial distribution

[Daniel Ezra Johnson]

Hi,

It sounded from Scott's question that the raters chose basically A or
B (re-analyzed or not). If so, is there a better alternative to
Cohen's kappa - I would have thought there probably was? (Sorry if I
misread it and there are more than two classes).

Dan

On Fri, Mar 26, 2010 at 11:58 AM, Maria Wolters <maria.wolters@ed.ac.uk> wrote:
> Hi everybody,
>
> the psy and psych packages in R have extensive implementations of kappa. If the definitions can be analysed into a small number of classes, then this might be sufficient.
>
> Cheers
>
> Maria
>
>
> Am 26.03.2010 um 15:02 schrieb Matt Goldrick:
>
>> Hi Scott,
>>
>>
>> I'm not precisely sure about your design, but if this is the core question:
>>
>> The student wants to know, essentially, to what extent the respondents
>> agree on the meanings.
>>
>>
>> Then I'd recommend using (generalized) Cohen's kappa--it provides an index of the degree to which multiple raters agree on their classification of categorical data.
>>
>> An online calculator can be found here (no endorsement of its accuracy!):
>>
>> http://cosmion.net/jeroen/software/kappa/
>>
>> hope that helps,
>> mG
>>
>> On Mon, Mar 22, 2010 at 1:29 PM, Scott F. Kiesling <kiesling@pitt.edu> wrote:
>> Hi linguistics stats gurus-
>>
>> I have a student who has response data from a questionnaire. Each
>> question presents a word and then asks to choose a definition. The
>> definitions are all categorized as reanalyzed or not reanalyzed.
>>
>> The student wants to know, essentially, to what extent the respondents
>> agree on the meanings.
>>
>> It seems that chi-square might be appropriate here, but there are
>> several responses per speaker, and her question is really about
>> speakers overall, I think.
>>
>> For wome reason I can't get my head around this one and I'm only an
>> amateur statistician anyway, so any suggestions are very much
>> appreciated.
>>
>> Best
>> Scott
>>
>> --
>> Scott F. Kiesling, PhD
>>
>> Associate Professor
>> Department of Linguistics
>> University of Pittsburgh, 2816 CL
>> Pittsburgh, PA 15260
>> http://www.linguistics.pitt.edu
>> Office: +1 412-624-5916
>>
>>
>>
>
>
> --
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
>
>