PubMed · 11349967
Comparison of methods to analyse imprecise faecal coliform count data from environmental samples.
Abstract
Imprecise values arise when bacterial colonies are too numerous to be counted or when no colonies grow at a specific dilution. Our objective was to show the usefulness of multiple imputation in analysing data containing imprecise values. We also indicate that interval censored regression, which is faster computationally in situations where it applies, can be used, providing similar estimates to imputation. We used bacteriological data from a large epidemiological study in daycare centres to illustrate this method and compared it to a standard method which uses single exact values for the imprecise data. The data consisted of numbers of FC on children's and educators' hands, from sandboxes and from playareas. In general, we found that multiple imputation and interval censored regression provided more conservative intervals than the standard method. The discrepancy in the results highlights both the importance of using a method that best captures the uncertainty in the data and how different conclusions might be drawn. This can be crucial for both researchers and those who are involved in formulating and regulating standards for bacteriological contamination.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
H Carabin, T W Gyorkos, L Joseph, P Payment, J C Soto. 2001. Comparison of methods to analyse imprecise faecal coliform count data from environmental samples.. https://doi.org/10.1017/s0950268801005222. https://pubmed.health/record/pubmed%3A11349967/comparison-of-methods-to-analyse-imprecise-faecal-coliform-count-data-from-environmental-samples
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite PubMed Health by Indiciae: BibTeX · RIS.