home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!news.gtech.com!noc.near.net!mars.caps.maine.edu!maine.maine.edu!cunyvm!psuvm!auvm!DGOGWDG1.BITNET!SERTEL
- Message-ID: <STAT-L%92111810423928@VM1.MCGILL.CA>
- Newsgroups: bit.listserv.stat-l
- Date: Wed, 18 Nov 1992 16:38:58 MEZ
- Sender: "STATISTICAL CONSULTING" <STAT-L@MCGILL1.BITNET>
- From: Suitbert Ertel <SERTEL@DGOGWDG1.BITNET>
- Subject: Clustered events
- Lines: 41
-
- Events occurring in clusters.
-
- Are there tools available determining degrees of event
- clustering across observational time periods? Is inferential
- statistical judgment possible?
-
- To illustrate: 20 events are recorded within a period of 40
- days. Maximum clustering would be given if the 20 events occurred on
- the same day. The other (=non-cluster) extreme is occurrence of
- 20 events exactly every two days. Both extremes deviate from
- random distribution, the first, clustering, is what I am
- interested in in the first place.
-
- Can you suggest, for this kind of data, a measure indicating
- degree of clustering? If possible, the indicator should be
- independent of numbers of events, i.e. it should allow for
- comparing, for a period of, say, 40 days, degree of clustering of
- A-events (say, N=20) with degree of clustering of B-events (say,
- N=60).
-
- Even more useful would be an indicator which, at the same time,
- were independent of the length of the observational period. That
- is, an observational period of, say 40 days might require
- comparison, regarding degree of event clustering, with an
- observational period of 50 days. But I could renounce that
- option for my present project.
-
- The presence of clustering for a one-sample case and differences
- among degrees of clustering for n>1 sample cases should be
- subjected to inferential statistical procedures.
-
- (At my disposal: STATISTICAL CALCULATOR by MOLE Software (Tony Dusoir)
- providing, among many others, routines regarding permutation
- and bootstrap functions. Programming (in C) possible.)
-
- Suitbert Ertel
- Institute of Psychology
- Gosslerstrasse 14
- 3400 Goettingen
- Germany
- e-mail: sertel@dgogwdg1.bitnet
-