Table 1.

Sample size calculation parameters

Za/2	Type I error	1.96
ZB	Type II error	0.84
λ 0	Estimated change in the accident rate per 1000 minibuses over the 14-month period in the control arm	−0.01
	Absolute difference	.01
λ 1	Estimated change in the accident rate per 1000 minibuses over the 14-month period in the intervention arm	−0.02
y	Number of person years in each arm	116
	Number of cars assuming cars followed for 1.16 years or 14 months	954
	Number of vehicles to follow with 10% buffer per arm	1060

The sample size calculation formula in Table 1 used is adopted from Hayes & Mouton (2009). A buffer of 10% was included to allow for buses that may drop out of the study during recruitment for any reason