Pilot study of statistical method of measuring changes in pesticide residues

Tuesday 22 May 2007

A way to measure changes in levels of pesticide residues in food over several years would be useful to assess the effectiveness of changes in industry practices to reduce residues. However, measuring such changes has proved difficult by conventional statistical means because of the nature of the residue data.

This includes factors such as the data being highly variable, many results being below the limit of determination, and the reporting limits vary, as does the range of pesticides tested.

Therefore, the Agency commissioned a pilot study using a novel statistical approach, which looked at the distribution of residue data (not just the mean values as in conventional approaches) and aimed to detect differences in the shapes of these distributions, over time and between domestic and imported crops. It used data on UK-grown apples, strawberries and protected lettuce in a range of years between 1998 and 2006.

Although the study results show some promise in detecting differences in the distributions of residues, this is still very limited by the nature of the residue data, particularly when most of the residue results are below the limit of determination.