In research, many studies are non-randomized, so risk of bias.

Newcastle-Ottawa scale is one attempt to assess bias formally – judged on:

• the selection of the study groups;
• the comparability of the groups;
• the ascertainment of either the exposure or outcome of interest for case-control or cohort studies respectively. 

So things like cohort not being representative; control group coming from different population; measurement being rather subjective; duration of follow up – all increase risk of bias.

Publication bias

Suspicious if small study with big effect!