Written by Using data from a project conducted with the University of California San Francisco and funded by the San Francisco Department of the Environment last year, Zabble (AI waste management app) sought to answer questions about the minimum required sample size for reliably estimating waste contamination.
The study was based on a 6-month experiment conducted at four university buildings (three medical research and one mixed-use), and was funded by the San Francisco Department of the Environment in collaboration with the University of California San Francisco (UCSF).
The experimental methodology involved training multiple sorters to objectively document contamination by sampling 10 to 15 landfill bags per day, which represented approximately 35% to 70% of the daily landfill-bound materials. Using this rich dataset, the researchers calculated the observed contamination levels based on smaller theoretical samples, which allowed them to establish statistical requirements for accurate data collection.
The statistical analysis focused on finding the point at which contamination levels stabilized, indicating that collecting more data did not significantly change the observed results. For overall building contamination, the data suggested that contamination levels level off as the number of bags sampled daily increases. Specifically, the difference in contamination becomes insignificant when it falls below 5%.
Across the four buildings tested, the experiment indicated that sampling only 4 bags per day for a month is sufficient to estimate a building’s contamination, meaning that sampling 5 or more bags per day yields an observed contamination difference of less than 5% compared to the 4-bag sample. When examining different time periods beyond one month, the analysis determined that the absolute minimum number of total samples required to estimate contamination (within the 5% difference threshold) is achieved with a sampling period of 1 week, at 8 bags per day.
To provide more granular insights crucial for sustainability and zero waste initiatives, the analysis was extended to include detailed information on specific contamination items. Researchers studied the prevalence of the top ten observed contaminants to understand how many bags needed to be sampled daily to obtain a reliable statistical understanding of these specific items. Similar to the analysis of overall contamination, the percentage of bags containing a specific top ten contaminant began to stabilize as the number of samples per day increased. To ensure the difference in the percentage of bags containing a top ten contaminant remained within 5%, the required sample size generally increased to 5 bags per day for a month.
These statistical findings directly inform best practices for zero waste initiatives and contamination monitoring. The conclusion emphasized two key statistical recommendations for optimizing sampling efforts: for a minimum number of samples needed to estimate contamination, a sample size of 1 week at 8 bags per day is required. However, to reliably estimate the prevalence of specific top contaminants, the sample size should be increased to 11 bags per day for a 1-week period.
You can read the full article at https://www.zabbleinc.com/blog-post/how-large-of-a-sample-size-do-you-need-to-estimate-contamination-accurately