Selection bias is an important consideration within research methodology. Selection bias has occurred when a person was selected for participation within research on a factor other than random sampling from within a sampling frame. People being chosen to participate in research because they are friendly, attractive, bored, etc. are all examples of selection bias. Alternatively different groups being recruited from different populations is also an example of selection bias. Considering only the evidence that confirms a hypothesis while ignoring contradictory evidence is known as cherry picking.
Religion and health research is often rife with selection bias issues. Research addressing religion often recruits a religious sample from faith-based organizations, but will recruit the non-religious sample in a different fashion. This approach is problematic because group membership in any social group (e.g., Baby Eaters Anonymous, the Roman Catholic Church, etc.) is associated with better health . Given that religious samples like these are from a social group, they would be expected to have a "health advantage" on that basis alone. For example, Frankel and Hewitt (1994) compared Christians to the non-religious and concluded that Christians were generally healthier. However, Christians were specifically recruited from faith-based clubs while the non-religious were recruited randomly from the community. The authors then brilliantly found that Christians were healthier. This would be like comparing IQ scores from random Christians to IQ scores from Satanists recruited from Mensa, and then concluding that being a Satanist was associated with higher intelligence.
In a different example, US-based faith-based initiative aimed at rehabilitating offenders, known as Innerchange Freedom Initiative, was found by researchers to be very effective at reducing re-offending rates.  However, their selection of which prisoners to include in their analysis was flawed due to their cherry picking of participants.
- "So, how did the Penn study get perverted into evidence that InnerChange worked? Through one of the oldest tricks in the book, one almost guaranteed to make a success of any program: counting the winners and ignoring the losers. [...] InnerChange started with 177 volunteer prisoners but only 75 of them "graduated." Graduation involved sticking with the program, not only in prison but after release. No one counted as a graduate, for example, unless he got a job. "
- Frankel, B. G., & Hewitt, H. E. (1994). Religion and well-being among Canadian university students: The role of faith groups on campus. Journal for the Scientific Study of Religion, 33, 62-73.
- Duwe G, King M., Can faith-based correctional programs work? An outcome evaluation of the innerchange freedom initiative in Minnesota, Int J Offender Ther Comp Criminol. 2013 Jul;57(7):813-41. doi: 10.1177/0306624X12439397. Epub 2012 Mar 21.
- Mark A.R. Kleiman, Faith-Based Fudging, Slate, Aug. 5 2003