# 3.2.2. Simulation approaches

Simulation or ‘Monte-Carlo’ methods (Manly, 1997) can be used to work out the best combination of “bees/cage” x “number of cages/group” given expected impact of a certain treatment. Given a certain average life span and standard deviation for bees of a control group and a certain effect of a treatment (in terms of percentage reduction of the life span of bees), one can simulate a population of virtual bees each with a given life span. Then a program can test the difference between the treated group and the control group using increasing numbers of bees (from 5 to 20) and increasing numbers of cages (from 3 to 10). The procedure can be repeated (e.g. 100 times) and a table produced with the percentage of times a significant difference was achieved using any combination of bees/cage x number of cages/group.

A program using a
*t*-test to determine these parameters
is given as online supplementary material. It is assumed that the dependent
variables in the bee population are normally distributed. The simulation can be
run another 100 times simply by moving the mouse from one cell to another.
Alternatively, automatic recalculation can be disabled in the excel
preferences.