Knowing the direction of the bias can be of practical importance to the investigator.

a) Suppose an investigator finds a very strong effect, say the estimated RR (in the study population) is 6, and s/he can also persuasively argue that any possible bias must be towards the null. Then what can be concluded about the correct (True) value of RR?

b) Suppose an investigator finds a very weak estimated RR (in the study population) of, say, 1.3, and can argue that any bias must be away from the null. Then what can be concluded about the correct (True) value of the RR