<P>This project applied two different statistical procedures to a data set containing a variety of consumption distributions, with a view to assessing their usefulness in estimation of food consumption patterns. This project did not consider contaminant levels in food but includes data on the consumption of Vitamin A (which is available in the Dietary and Nutritional Survey of British adults (1990)).</P>
<p>The classical approach to estimate the within-individual and the between-individual components of variation separately is the analysis of variance (ANOVA) method used by Beaton. A more sophisticated procedure by Nusser can manage different types of statistical variability, intake values and between-person intake variability ranges and is able to process an array of distributions within the data. These two models were applied using data from different food groups: those whose consumption is fairly evenly distributed; those where it is slightly skewed; and those where consumption is very skewed. The project also considered whether 4-day food diary data could be used effectively rather than 7-day data. </P>
<p>When there is a risk of chemical contaminants being found in food, e.g. mycotoxins or pesticide residues, it is important to know how frequently and in what amounts a person might be consuming that food. There is a need to develop a model that can estimate long-term intake from short-term data available, with a view for assessment of when it becomes likely that individuals may consume contaminants to a potentially hazardous level. </P>
<p>Find more about this project and other FSA food safety-related projects at the <a href="http://www.food.gov.uk/science/research/" target="_blank">Food Standards Agency Research webpage</a>.