Growth curve models of repeated binary response.

Area of Science:

• Biology
• Statistics
• Biostatistics

Background:

• Repeated measures of binary responses over time are common in biological studies.
• Characterizing response patterns over time is crucial for understanding biological processes.
• Continuous response variables are often analyzed using growth curve models like Potthoff-Roy.

Purpose of the Study:

• To adapt growth curve modeling strategies for binary response data.
• To implement a weighted least squares (WLS) method for analyzing repeated binary outcomes.
• To demonstrate a method for characterizing temporal response patterns in biological experiments.

Main Methods:

• Utilized weighted least squares (WLS) methods for binary response data.
• Constructed growth models using polynomial functions on marginal response.
• Developed a strategy to drop nonsignificant higher-order polynomial functions to avoid small-sample problems.

Main Results:

• Successfully implemented a growth curve modeling strategy for repeated binary data.
• The proposed WLS method effectively models response patterns over time.
• Demonstrated the model's utility with an example of fly oviposition behavior.

Conclusions:

• Growth curve modeling can be effectively applied to repeated binary response data using WLS.
• The method provides a robust approach for analyzing biological data with temporal dependencies.
• This technique aids in understanding biological responses over time, particularly in scenarios with limited sample sizes.