The investigation of samples containing many species: II. Sample comparison
A comparison is made between a number of indices of similarity fer describing the similarities between multiple species samples, namely K, I, (1 ‐ z), (x2/AT)1/2, r and a weighted similarity coefficient under various transformations. After discussing theoretical aspects of the indices, they are test...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Published: |
Oxford University Press
1971
|
| Subjects: | |
| Online Access: | http://eprints.um.edu.my/24497/ https://doi.org/10.1111/j.1095-8312.1971.tb00171.x |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | A comparison is made between a number of indices of similarity fer describing the similarities between multiple species samples, namely K, I, (1 ‐ z), (x2/AT)1/2, r and a weighted similarity coefficient under various transformations. After discussing theoretical aspects of the indices, they are tested on two sets of data obtained by mist‐netting birds. The resulting matrices are sorted by a simple nearest‐neighbour agglomerative technique and later by principal component analysis. It is concluded that the V index, which is Kendall's rank correlation coefficient (Kendall, 1955), provides the most meaningful segregation, particularly under principal component analysis. Problems of subjective interpretation are discussed. Copyright © 1971, Wiley Blackwell. All rights reserved |
|---|