You are here:
Using Ingredient Lists to Quantitatively Characterize Composition of Consumer Products
Citation:
Price, P., K. Phillips, AND K. Isaacs. Using Ingredient Lists to Quantitatively Characterize Composition of Consumer Products. ISES 2015 Annual Meeting, Henderson, NV, October 18 - 22, 2015.
Impact/Purpose:
The poster provides a new methodology of deriving estimates of weight fractions
Description:
Assessing exposure to substances in consumer products requires data on the composition of the products. This is a challenge since product composition data are rarely available. Many products, however, provide a list of ingredients. In many cases the list is presented in descending order by weight fraction and includes all ingredients above some cut off weight fraction. We have developed a probabilistic model to estimate the range of weight fractions for an ingredient as a function of the number of reported ingredients for a product (list length) and the place of the ingredient on an ingredient list (rank). This is possible because the range of weight fractions for an ingredient is constrained by the product’s list length, the ingredient’s rank, and the weight fraction of the unreported ingredients. The model provides the medium and 95% one tail upper confidence limit (UCL) of the range of weight fractions. Weight fractions predicted by the model are compared to composition data obtained from a database of Material Safety Data Sheets. The model suggests that estimates of MLE and 95th UCL are relatively insensitive to assumptions about the shapes of the distributions of weight fractions for ingredients of a specific rank across products or the size of the total weight fraction of unreported ingredients. The values produced by the model may provide useful screening values for characterizing the composition of products with ingredient lists.
