A survey of laboratory and statistical issues related to farmworker exposure studies

Developing internally valid, and perhaps generalizable, farmworker exposure studies is a complex process that involves many statistical and laboratory considerations. Statistics are an integral component of each study beginning with the design stage and continuing to the final data analysis and interpretation. Similarly, data quality plays a significant role in the overall value of the study. Data quality can be derived from several experimental parameters including statistical design of the study and quality of environmental and biological analytical measurements. We discuss statistical and analytic issues that should be addressed in every farmworker study. These issues include study design and sample size determination, analytical methods and quality control and assurance, treatment of missing data or data below the method's limits of detection, and post-hoc analyses of data from multiple studies.