A method for testing association patterns of social animals

Area of Science:

• Ecology
• Behavioral Ecology
• Conservation Biology

Background:

• Association indices are used to measure social connections between individuals.
• Existing methods lack statistical validation for comparing observed associations to random expectations.
• This gap hinders accurate interpretation of social structures in animal populations.

Purpose of the Study:

• To adapt a statistical test for validating association index values against random associations.
• To introduce a method that retains key ecological data features during randomization.
• To assess social association patterns in Hector's dolphin (Cephalorhynchus hectori) populations.

Main Methods:

• Adaptation of Manly's (1995) randomization test for association data.
• Computer-generated randomization preserving original data characteristics (e.g., group size, sighting frequency).
• Application to four datasets from two Hector's dolphin populations using the Half-Weight Index (HWI).

Main Results:

• The new test identified significant differences in association patterns between Hector's dolphin populations.
• Populations with similar median Half-Weight Index values exhibited distinct social structures (random vs. non-random associations).
• The method effectively differentiates between random and non-random social associations.

Conclusions:

• The developed statistical test is crucial for validating the analysis of association indices.
• This approach provides a robust method for ecological and behavioral studies of sociality.
• Accurate assessment of social structures is vital for understanding population dynamics and conservation efforts.