Assessing Sample Number Requirements for Municipal Solid Waste Composition Studies

Abstract

Solid waste decision makers use waste composition studies in sustainability planning and as metrics to measure their environmental benefits (e.g., recycling rates). In many of these studies, there are varying waste components and number of samples assessed. Often, studies will rarely determine whether the number of samples sorted was actually sufficient to be statistically representative. We collected 24 US studies and assessed whether their material categories were statistically representative at a 90% confidence level and 10% precision following current standards. For each study and each material category, we calculated the number of samples required to be representative and compared it with the number of samples conducted by the study. Of the 24 studies, 17 had one to seven representative material categories, where 12 had one or two representative materials out of the average 44 original materials assessed per study. Nonrigid plastic film and food waste were most frequently statistically representative. Each study's original category list was condensed into 12 categories, and the analysis was repeated. Two studies that originally had no categories representative now had at least one. With the condensed materials, the mixed paper, mixed plastic, and food categories were most representative. The findings can be used by decision makers to gain a more comprehensive understanding of the importance of data quality analysis since most materials were not statistically representative; two approaches to being more representative are to sort more samples or condense the material list.