An evaluation of sampling methods used to produce insect growth models for postmortem interval estimation

Area of Science:

• Forensic Entomology
• Developmental Biology
• Statistical Modeling

Background:

• Carrion insect development data aids in estimating postmortem intervals.
• Existing literature lacks standardized protocols and investigates various sampling methods.
• The impact of sampling methods on predictive model performance remains unexamined.

Purpose of the Study:

• To investigate how different sampling methods influence predictive models for forensic entomology.
• To assess the reliability of age estimation from carrion flies using various subsampling techniques.
• To provide recommendations for optimal data collection in forensic insect development studies.

Main Methods:

• Calculated 95% inverse prediction confidence limits for growth curves of Chrysomya megacephala and Sarconesia chlorogaster.
• Utilized confidence limits from entire age cohorts as a benchmark for realism.
• Evaluated the performance of predictive models derived from different subsampling schemes.

Main Results:

• Random subsamples produced predictive models comparable to those from complete data.
• Subsampling the largest larvae resulted in poor model performance with overly narrow confidence intervals.
• Models based on entire cohorts provided the most realistic confidence limits for age estimation.

Conclusions:

• Measuring entire age cohorts is recommended for forensic insect development studies.
• Subsampling strategies, particularly focusing on largest larvae, can lead to inaccurate minimum postmortem interval estimations.
• Standardized protocols are needed to ensure the reliability of forensic entomology data.