Pawel,
Thanks ever so much, that definitely helps!
Having said that, it would be great if you could send me an electronical version of the paper, considering the photocopy I got hold of was of so poor quality.
Also, I gather you have a certain familiarity with the topic, so may I ask in passing what you think about Arnold's test? I am a bit concerned that it might be out dated, but I find it conceptually easy to understand and share with non-statisticians.
Thanks again for your help
Kind Regards
Sandra

Sandra Alba
University Medicine - Level 7
Derriford Hospital
Plymouth
PL6 8DH
UK

> -----Original Message-----
> From: Classification, clustering, and phylogeny estimation [SMTP:[log in to unmask]] On Behalf Of Pawel P. Baran
> Sent: 16 March 2004 13:56
> To:   [log in to unmask]
> Subject:      Re: Arnold's Test for Clusters
>
> > > I have a not very statistical query regarding the paper "A test for
> Clusters" published in 1979 (Journal of Marketing Research, vol XVI,
> november 1979, 545-51). I would like to apply Arnold's test to my data. I
> have finally got hold of a copy of the paper, but it is the fudgiest paper
> ever! Typically, the most incomprehensible part is the end of page 548,
> where the two key functions are given. This is what I can recover (?
> indicate values which I cannot read):
> > >
> > > C(unimodal)= [e^(0.?6???)*g^(0.???42)*p^(0.21011)*a^(0.13??)]/N^(0.2972)
> > > C(uniform)=[e^(0.?????)*g^(0.89510)*p^(0.138??)*a^(0.163??)]/N^(0.236)
>
> I hope this would help:
>
> C(unimodal)= [e^(0.06239)*g^(0.95242)*p^(0.21011)*a^(0.13389)]/N^(0.29723)
> C(uniform)=[e^(0.08664)*g^(0.89510)*p^(0.13896)*a^(0.16356)]/N^(0.23672)
>
> Please let me know if you want the whole paper emailed (~700kb).
>
> Pawel
>
> --
> This email has been electronically scanned using
> DynaComm i:mail filtering software.
>
> If you have any questions please email
> [log in to unmask]
>
> Tony Daniel
> Sam Shayer
> Information Security Team
>