How many to sample? Statistical guidelines for monitoring animal welfare outcomes - Fig 5

Heat maps illustrating the relationship between the probability of an adverse event (x-axis), the scaling factor to specified value (y-axis), and the required sample size (yellow-grey gradient) for different Type I error rates (1–20%; a–d). For a given probability of an adverse event, larger sample sizes are required with low Type I error rates and at low scaling factors. The sample size is given up to a maximum of 10,240; larger sample sizes are required for the grey-coloured zone.