Design and analysis of multiple-choice feeding-preference experiments.

Area of Science:

• Ecology
• Marine Biology
• Statistical Methodology

Background:

• Multiple-choice feeding preference experiments are crucial in ecology but lack rigorous statistical analysis.
• Existing methods often fail to account for the interdependence of food types and autogenic changes.
• A majority of published studies in the littoral marine context have used incorrect analytical approaches.

Purpose of the Study:

• To propose a rigorous statistical method for analyzing multiple-choice feeding-preference experiments.
• To address the challenges of non-independent food offerings and autogenic changes in experimental arenas.
• To provide both parametric and nonparametric procedures for robust data analysis.

Main Methods:

• Utilized multivariate statistical analysis to handle the lack of independence between simultaneously offered food types.
• Employed basic statistical theory to account for autogenic changes within each food type using control arenas.
• Applied the proposed method to analyze feeding data of the sea urchin Tetrapygus niger offered three algae species.

Main Results:

• The new parametric and nonparametric methods yielded consistent results.
• Analysis revealed that Tetrapygus niger exhibits non-random feeding behavior, preferring Ulva nematoidea over other algae.
• The method requires an equal number of replicates in treatment and control arenas, exceeding the number of food types.

Conclusions:

• A rigorous statistical framework is now available for multiple-choice feeding-preference experiments.
• The proposed method accurately identifies specific feeding preferences, crucial for understanding nutritional biology and ecosystem dynamics.
• This statistical approach is applicable to various multiple-choice experimental designs beyond feeding studies.