Design and Analysis of Chronic Aquatic Tests of Toxicity.
Final rept. 15 Sep 79-31 Oct 80,
OHIO STATE UNIV RESEARCH FOUNDATION COLUMBUS
Pagination or Media Count:
The present investigation considers aspects of the planning of, reporting results from, and the statistical analysis of data arising from chronic aquatic tests of toxicity with fathead minnows. The discussion emphasizes analyses and data displays for the interpretation of qualitative mortality and abnormality data and quantitative weight data. The report consists of nineteen sections. Section I-IV describe the literature search, argue for and present a set of standardized data reporting formats, and contain some suggestions for further improving and standardizing the test procedure. Section V briefly describes the early life stage testing procedure and the resulting data. Sections VI-XIX discuss various topics pertaining to the statistical analysis of such data and to statistical aspects in the design of toxicity tests. Techniques discussed include preliminary graphical displays, tests for homogeneity among tanks within treatment groups, adjustments to account for tank heterogeneity within groups, outlier detection procedures, overall tests for treatment effects, multiple comparions and pairwise confidence intervals of treatment groups with the control group, analysis of variance, regression analyis and parametric and nonparametric dose response curve fits along with associated estimates of safe concentrations. All suggested procedures are illustrated with real examples based on aquatic toxicity data. Several newly developed, specialized computer programs are described to carry out nonstandard analyses. Author
- Water Pollution and Control