Table 2. Overview of the default and nine alternative designs used to investigate the effect of autocorrelation on power.

Situation	# of possible start moments, k	correlated B1 and I2 Mean, rBI	Equal # of obs. in B and I, #obsB = #obsI	Non-overlapping k	number of measurements, t	Mean diff. B and I, d	sd B	sd I
default	3	FALSE	TRUE	TRUE	60	1	1	1
k = 2	2	FALSE	TRUE	TRUE	60	1	1	1
k = 4	4	FALSE	TRUE	TRUE	60	1	1	1
rBI	3	TRUE	TRUE	TRUE	60	1	1	1
#obsB = #obsI	3	FALSE	FALSE	TRUE	60	1	1	1
Non Overlap k	3	FALSE	TRUE	FALSE	60	1	1	1
d = .3	3	FALSE	TRUE	TRUE	60	0.3	1	1
d = .6	3	FALSE	TRUE	TRUE	60	0.6	1	1
sdB = 2sdI	3	FALSE	TRUE	TRUE	60	1	1.33	0.67
2sdB = sdI	3	FALSE	TRUE	TRUE	60	1	0.67	1.33

Shaded cells show the factor level that is different from the default scenario.

¹B = Baseline,

²I = Intervention