Estimation of growth parameters using a nonlinear mixed Gompertz model.

Area of Science:

• Animal Science
• Biostatistics
• Quantitative Biology

Background:

• Accurate estimation of growth parameters and variances is essential for effective decision-making using simulation models.
• Traditional models may not adequately account for individual variations and correlations in longitudinal data.

Purpose of the Study:

• To compare the utility of a mixed Gompertz growth model versus a fixed effects model for analyzing longitudinal growth data.
• To assess the impact of accounting for between-bird variation and heterogeneous variance on growth parameter estimation.

Main Methods:

• Application of a mixed Gompertz growth model to longitudinal body weight (BW) data.
• Comparison of variance partitioning (between- and within-bird) and residual variance reduction between mixed and fixed effects models.
• Evaluation of bias reduction in estimation due to selective sampling.

Main Results:

• The mixed model effectively partitioned BW variation into between- and within-bird components.
• A significant decrease (over 55%) in residual variance was observed with the mixed model.
• The mixed model reduced estimation biases arising from selective sampling.
• The covariance structure in the mixed model accounted for BW correlations within individuals across ages.

Conclusions:

• Mixed effects growth models are recommended for the analysis of longitudinal growth data.
• These models offer superior accuracy in parameter estimation and variance partitioning compared to fixed effects models.
• The use of mixed models enhances the reliability of simulation models for informed decision-making in biological studies.