Population Size Estimation using Zero-truncated Poisson Regression with Measurement Error.

Area of Science:

• Ecology
• Population Biology
• Biostatistics

Background:

• Population size estimation is crucial in biological sciences.
• Covariates measured during capture can have temporal variations, introducing measurement error.
• Ignoring covariate measurement error can lead to underestimation of population size.

Purpose of the Study:

• To address population size estimation challenges with covariate measurement error.
• To develop and validate methods for correcting measurement error in biological data.
• To improve the accuracy of population size estimates in capture-recapture studies.

Main Methods:

• Developed a conditional score approach for population size estimation.
• Introduced a nonparametric corrected score approach.
• Evaluated estimator performance through extensive simulations and real-world data analysis.

Main Results:

• A naive estimator underestimates population size when measurement error is present and normally distributed.
• The proposed conditional score and nonparametric corrected score approaches provide consistent population size estimates.
• The nonparametric approach does not require normality assumptions for covariates or measurement errors.

Conclusions:

• Accurate population size estimation requires accounting for covariate measurement error.
• The developed methods offer robust solutions, particularly when covariate distributions are unknown or non-normal.
• These methods enhance the reliability of population estimates in ecological research.