Statistical analysis of the cladoceran reproductivity test
Martin A. Hamilton · Environmental Toxicology and Chemistry · 1986
Abstract The data from a renewal toxicity test are awkward to interpret if some of the adult organisms die during the experiment. The usual summary statistic is the mean number of young produced per adult, but when some adults die, the mean can be defined in two different ways, the mean overall (MOA) and the mean ignoring mortality (MIM). The MOA will be small if adults die but the MIM is relatively unaffected by mortalities. The MIM estimate is valid only if the effect of the toxicant on survival is not correlated with the effect on production of young. Analysis of data from several 7-d cladoceran renewal tests indicates that this correlation is negligible. For such tests, it appears that MIM calculations provide a valid way to separate the effect of the toxicant on the production of young from the effect on the survival of the adults. For both MOA and MIM, methods of statistical analysis are presented, including methods of point estimation, standard error estimation, interval estimation and significance test calculation.