Pooling Data for Pest Management Analysis: A Stein-rule Approach

摘要

The study proposes a method to enrich prospective Integrated Pest Management (IPM) data sets by pooling sample data originating from different experimental designs. Predictive risk performance of alternative estimation procedures are compared, and a Stein-rule estimator is proposed as an alternative to traditional statistical approaches to pooling data. The Stein-rule estimator adjusts between ordinary least squares estimates and restricted least squares estimates based on the correctness of the restrictions. This type of analysis is required for any investigation of economic thresholds because the procedure will indicate whether the various restrictions associated with pooling conform to the data sets.