Selection Bias: Statistical Integrity in Christian Community
One day a number of concerned mothers met with the minister to express their frustration and anger over the unseemly conduct of a particular boy in Sunday School. They did not want their children exposed to this child and feared what he represented. For it seemed that this boy was modeling “bad behavior” – verbal outbursts that sometimes involved profanity, a lack of sensitivity to other children’s personal space (occasionally biting them when irritated or provoked) and an unpredictably violent imagination when playing with toys. No Sunday school is equipped to handle problems of this magnitude. So upon expressing their indignation, the mothers requested that the minister call the child’s parents and ask that he not return to Sunday school. Obviously, there were family issues that needed serious and immediate attention.
The “problem child” was ours. My wife received the call early one morning. The minister was deeply apologetic and pastoral in his approach. But the damage had been done. What were we to do? Where could we go? Over the years, we had been through behavioral programs, family counseling, and psychiatric care. At this point, we were just beginning to come to terms with our son’s recent diagnosis: Tourette’s syndrome. Later, he would also be diagnosed with Asperger’s syndrome, bipolar disorder, and obsessive-compulsive disorder. But at this point he was about seven years old, and we knew only of the Tourette’s. We stopped attending this church. In fact, we stopped attending church altogether. — Thomas E. Reynolds, Vulnerable Community: A Theology of Disability and Hospitality
Engineering does not often apply directly to faith, but one method that has transformed the way I view community is a commitment to statistical honesty. In reading papers and critiques of clinical trials, one thing that comes up repeatedly is the question, “Is the community they engaged in this trial one that is diverse? Does it represent society in general? Can it translate into meaningful implications for the people I treat? Or were these participants selected in a biased way to favor a certain outcome? Is there a skew that limits how we may interpret and understand the world?”
One day it struck me to think about my own community with a similar critique. If I took a random sample of my friends from work, my neighborhood, and my church, would it look like it was truly random? Would there be an overrepresentation of certain types of people or a paucity in others? Would that statistical bias be a reflection of intentionality or a revelation in exclusivity?
I did a brief mental estimation and was not happy with the results. It is my natural human tendency to surround myself with others who think like me, talk like me, and act like me. What I have been grateful for in the work of medicine is being forced into contact with those who are very different from me, those whom, I am ashamed to say, I would not ordinarily choose as neighbors, associates, or friends. Through this means of grace, in the past year alone I have encountered former drug dealers and drug addicts, millionaires and mansion owners, wheelchair riders and deaf academics, judges and janitors, Holocaust survivors and pedophiles, saints and sinners. Though my coworkers (and myself) have often varied in expressions of compassion, we were obligated by both law and ethic to work with them in seeking their greatest benefit.
And so I found myself wondering, “Who is my neighbor? And have I shaped the courses of my encounters, friendships, and associations to suit their needs or my own?” I found that I did not like the answer: that my friends were mainly from certain ethnic groups, certain socioeconomic demographics, certain intellectual capacities and predispositions, certain persuasions of personality and even certain sects of faith. I had groomed and self-selected myself into becoming a statistical outlier in ways incompatible with the gospel, and it grieved me to think of those I had hurt in my exclusivity.
In this season of Lent, it is both sobering and encouraging to consider Christ’s disabled state, the divinity of he whose statistical cross-section of acquaintances included fishermen and Pharisees, tax collectors and political zealots, Samaritans and the blind, lepers and the governor’s wife, Centurions and servants:
Who, being in very nature God,
did not consider equality with God something to be grasped,
but made himself nothing,
taking the very nature of a servant,
being made in human likeness.
And being found in appearance as a man,
he humbled himself
and became obedient to death—
even death on a cross!