|
|
|
|
CONTRIBUTORS:
|
|
|
INSTITUTION ID:
|
|
|
SERIES TITLE:
|
|
|
YEAR:
|
2
|
|
PUB TYPE:
|
Working Paper/Manuscript
|
|
WORKING PAPER NUMBER:
|
None
|
|
PAGES:
|
|
|
SUBJECT(S):
|
Log odds-ratio; Yates’s correction; exact test; contingency table
|
|
DISCIPLINE:
|
Ecology
|
|
HTTP:
|
|
|
LANGUAGE:
|
English
|
|
PUB ID:
|
103-394-492
(Last edited on
2005/06/15 11:48:00 GMT-6)
|
|
SPONSOR(S):
|
|
|
ABSTRACT:
Goodman’s log odds-ratio is introduced to test interspecific association. As to association analysis, the sensitivity of the loglinear method is between that of often-used Pearson χ2 with Yates’s correction and that of Gilbert & Wells’s exact test (log-likelihood ratio). Results of log-linear method and exact test are close, and both can exactly indicate the association relationship of some extreme examples. In the field study, exact test will overdraw the number of positive associated species pairs. Moreover, the new method can show the polarity of association directly, and its significant test is very convenient. Therefore, we believe that the method should have a wide application in species association analysis.
|
|
|
|
STATISTICS
|
|
Click on # to view
|
|
Citations
|
|
0
|
|
References
|
|
2
|
|
Comments
|
|
0
|
|
Quality
|
|
0/0.00
|
|
Interest
|
|
0/0.00
|
|
View(er)s
|
|
2/306
|
|
|
|
|
|
|
| Prev |
Next |
|
