ABSTRACT
Precise Estimate of LC50 ............................................................................80 5.5 When a Time to Event Model Might be Less Costly
than a Traditional Experiment...................................................................82 5.6 Conclusions...................................................................................................82 Acknowledgments ................................................................................................83 Appendix................................................................................................................83 References...............................................................................................................87
In a traditional fixed-time study of toxicant effects, the number of alive and dead individuals in each experimental condition is recorded at a fixed time (e.g., 96 h). If the purpose is to estimate LC50, the median lethal concentration, the experimental conditions are different concentrations. If the purpose
is to evaluate individual or environmental characteristics that affect sensitivity to a toxicant, the experimental conditions are groups with different individual characteristics or different exposure environments. Estimates of LC50 or treatment effects are based on the proportion of dead individuals in each experimental condition. Many different techniques, including probit analysis, logistic regression, the trimmed Spearman-Karber approach, and contingency table analysis can be used to estimate dose-response curves, LC50s, or treated effects. These techniques are described in many textbooks;
Hamilton
