Statistical Design
The number of animal subjects required for an investigation will depend on the questions being explored. Field studies and laboratory studies typically require greatly different statistical designs, with field studies typically requiring much larger numbers. The life stage of the fish used in each study will also affect the numbers needed. Studies of early life stages typically require very large numbers of individuals. In all cases, studies should be designed to use the fewest animals necessary to reliably answer the questions posed.
The use of adequate numbers to establish variance and to assure reliability is essential so as to prevent needless repetition of the study (ASIH et al. 1987, 1988). A true "replicate" is the smallest experimental unit to which a treatment can be applied independently.
Pseudoreplication can result from wrongly treating multiple samples from one experimental unit as multiple experimental units or from using experimental units that are not statistically independent (Heffner et al. 1996). Study objectives should be presented as clearly stated hypotheses, and explanations should be provided as to the need for the type and quantity of data to be collected as well as what will constitute an end to the experiment. Power analysis procedures have been useful to many researchers to determine the appropriate number of fish needed to accomplish acceptable, statistically valid results. Researchers are encouraged to consult with a statistician to develop study designs that have the appropriate statistical power to accomplish research objectives.
Mortality as an Experimental Endpoint
In general, experimental endpoints other than death of the experimental subjects should
be developed unless death is required by the study protocol. The use of mortality as an
endpoint is appropriate when one or both of the following criteria are met:
(1) Little or no information pertaining to research objectives is available on the species of interest or the experimental variable being imposed. (For example, shortterm, limited mortality studies may be used to develop experimental limits for subsequent sublethal studies.)
(2) Mortality data are required or at least used frequently by a sponsoring agency to provide a basis for criteria development as part of a regulatory process. Many studies concerning the effects of pathogens and parasites or studies concerning the effects of drugs and other chemicals require mortality endpoints.
The number of animal subjects required for an investigation will depend on the questions being explored. Field studies and laboratory studies typically require greatly different statistical designs, with field studies typically requiring much larger numbers. The life stage of the fish used in each study will also affect the numbers needed. Studies of early life stages typically require very large numbers of individuals. In all cases, studies should be designed to use the fewest animals necessary to reliably answer the questions posed.
The use of adequate numbers to establish variance and to assure reliability is essential so as to prevent needless repetition of the study (ASIH et al. 1987, 1988). A true "replicate" is the smallest experimental unit to which a treatment can be applied independently.
Pseudoreplication can result from wrongly treating multiple samples from one experimental unit as multiple experimental units or from using experimental units that are not statistically independent (Heffner et al. 1996). Study objectives should be presented as clearly stated hypotheses, and explanations should be provided as to the need for the type and quantity of data to be collected as well as what will constitute an end to the experiment. Power analysis procedures have been useful to many researchers to determine the appropriate number of fish needed to accomplish acceptable, statistically valid results. Researchers are encouraged to consult with a statistician to develop study designs that have the appropriate statistical power to accomplish research objectives.
Mortality as an Experimental Endpoint
In general, experimental endpoints other than death of the experimental subjects should
be developed unless death is required by the study protocol. The use of mortality as an
endpoint is appropriate when one or both of the following criteria are met:
(1) Little or no information pertaining to research objectives is available on the species of interest or the experimental variable being imposed. (For example, shortterm, limited mortality studies may be used to develop experimental limits for subsequent sublethal studies.)
(2) Mortality data are required or at least used frequently by a sponsoring agency to provide a basis for criteria development as part of a regulatory process. Many studies concerning the effects of pathogens and parasites or studies concerning the effects of drugs and other chemicals require mortality endpoints.