# 35. Statistical manipulation - Biased Sampling - Selective bias ## 35.5. Methodology/Refinements/Sub-species ### 35.5.4. Selecting favourable criteria This is a deliberate or accidental bias in the criteria selected to determine if a particular sample has been affected by a particular cause. For example, in a drugs trial it is possible to choose a wide range of possible dependent variables to establish whether a drug is useful, usable, and safe under a lot of differing conditions. Choosing favourable criteria can alter the apparent outcome of a trial. For instance, if a drug works well for a short time like 2 years but becomes useless after 2 years of use, a trial could seek to use favourable criteria like "Effectiveness at 3 months, 6 months, 12 months and 24 months" and ignore results after that. The study would have conformed to its own terms of reference, would show excellent results for effectiveness, but it would not show the true picture because the criteria have been manipulated to suit a hidden agenda.