Hello
Maurizio Pisati wrote: 2) Since each province has a different population, I guess I should
standardized the cell frequencies in such a way to get rates (e.g., number
of units of a given NRM per 100.000 population). Again, what are the best
analytical techniques to apply?
A statistical technic called Decentred Correspondence Analysis (DCOA) in the ADE4 documentation -
see COA module documentation - allows to run a COA with a table containing rates. The denominator
must be the same for a given row as it is the case in your data with the population of each
province. You will find a good description (in french) of DCOA in "Doledec,S., Chessel, D.,
Olivier, J.M., (1995) L'analyse des correspondances decentrée: application aux peuplements
ichtyologiques du Haut Rhône. Bull. Fr. Pêche Piscic., 336, 29-40. This technic has also been used
recently in animal epidemiology to describe associations and competitions between different lung
infectious agents in Sheep in Senegal (See Tillard E., Faye B., 1997, Analyse d'un tableau de
taux: exploration multidimensionnelle de la circulation des agents infectieux pneumotropes chez
les petits ruminants au Sénégal, Topic documentation in ADE4 web site, volume 5 K-tables, analyse
d'un tableau de taux). I will send you an abstract of this documentation in english.
Regards
Emmanuel TILLARD
CIRAD-Elevage
Ligne Paradis
97410 St Pierre
La Réunion
Tél: 0 262 35 76 53
0 262 35 76 50 (standard)
Fax: 0 262 35 76 51
Please use E-Mail address: tillard@paradis.cirad.fr
This archive was generated by hypermail 2b30 : Sat Feb 10 2001 - 10:21:40 MET