Dear Colleagues
I am finishing a publication (for Acta Chiropterologica) on bat feeding
strategies, using contingency tables and log-linear analysis to explain the
effects of habitat and seasonality on bat preferences for particular insect
preys. I need to explain the reason why collapsing rows and/or columns, in
other words reducing the number of categories for a variable, (i.e. collapsing
months into wet and dry seasons) could decrease statistical power if this will
actually increase sample size within the table cells. Thanks for any assitance
or reference you could provide me to solve this question.
Pablo
Pablo Jarrin
Department of Biology
Boston University
|