Table 3.

Major points in 57 methods publications addressing allocation bias directionality and mechanisms

Direction of bias	Na (%)
Studies explicitly commenting on direction of bias	27 of 57 (47 %)
Studies commenting only on exaggeration of treatment effect	16 (28 %)
Studies commenting on possible bias in both directions	11 (19 %)
Studies not explicitly commenting on direction of bias	30 of 57 (53)
Mechanisms of bias
Allocation concealment protects against allocation bias	53 of 57 (93 %)
Preferential allocation of patients may lead to bias.	32 of 57 (56 %)
Allocation bias is possible in trials with restricted randomisationb, if upcoming assignments can be guessed.	11 of 57 (19 %)

^aNumber; ^brestricted randomisation refers to the process of making restrictions to the randomisation scheme (e.g. blocked randomisation)