Wednesday, October 07, 2015

Wolves, sheep, and sheepdogs


A follow-up post to this:


Among other things, pastors have a duty to protect the flock. That's not a question of forgiveness, but wisdom. A prudential question.

In risk-assessment you have paired factors:

i) What are the odds that x will happen?

ii) What's the harm if x happens?

Take this anecdote:

He became even more of a distraction when he produced equations that showed the possibility that a fission weapon could ignite the world's atmosphere. It was later discovered his calculations were wrong -- and a dozen other men made similar mistakes later -- but work stopped until the flaw was found. 
http://news.stanford.edu/news/2003/september24/tellerobit-924.html

Suppose there'd been a tiny risk of that eventuality? Even so, if global annihilation had been a live possibility, that's an unacceptable risk, even if the chances are very low. 

Likewise,

i) What's the risk of doing x?

ii) What's the risk of not doing x? 

Take vaccination. That carries a low risk of harm. But it's generally riskier not to vaccinate. As a rule, not to vaccinate does more harm. The benefit outweighs the risk. 

7 comments:

  1. "The benefit outweighs the risk". True. But then '[the] sluggard says, “There’s a lion in the road, a fierce lion roaming the streets!”' (Prov 26:13). "Benefit" and "risk" are subjective evaluations, which means that they may differ between people depending on their priorities.

    Since the Wilson case is indirectly under discussion, let me use it as an example. I haven't investigated in fine detail, but I believe all the following are true:
    (1) the man in question was guilty of child molestation and jailed
    (2) after he was paroled, he wished to get married
    (3) most (but not all) of those with supervisory responsibility agreed that this was a net good
    (4) he and his wife had a child
    (5) he is now accused of molesting his own child

    The critical issue to note is #5 - the alleged victim is his own child, not anyone else's. I am not aware of any report suggesting a third party has been put at risk, except for whatever distress the situation might be causing his wife.

    Speculate that the accusation is true, and that appropriate responses are applied. Now further consider that the real "risk / reward" calculation is not a third party child being molested, but his own child's existence. The tradeoff is not "restriction on freedom" vs "injury", it's "not existing" vs "injury". Those who argue that Wilson *and the parole board* are clearly in the wrong are in fact claiming that it would have been better that the child never be born than that it was put at risk of its father re-offending. That's a legitimate position to take, but it rather raises the bar on the discussion and on claims of recklessness.

    A similar argument applies to vaccination. How does a "high chance" of autism trade off against a "low chance" of measles? What about the risk of being involved in a car accident versus never leaving the house? Humans are notoriously bad at judging absolute risks and at comparing relative risks. What may be an obvious trade-off to one party may be a foolish trade-off to another. And if so, is it because the risks themselves are being mis-evaluated, or because priorities differ? There might be many layers of discussion involved to come to an agreement on highly contradictory "obvious" positions.

    ReplyDelete
    Replies
    1. "'Benefit' and 'risk' are subjective evaluations, which means that they may differ between people depending on their priorities."

      And the priority should be to protect the innocent.

      "Since the Wilson case is indirectly under discussion, let me use it as an example."

      You're free to do so. However, I initially brought up the Wilson case as a launchpad to discuss the issue of pastoral responsibilities more generally in regard to potentially dangerous parishioners.

      "The critical issue to note is #5 - the alleged victim is his own child, not anyone else's. I am not aware of any report suggesting a third party has been put at risk."

      Pedophiles pose a threat to any child who's a target of opportunity. Turn your back on the child for a few moments, and when you look back, it maybe too late. Take child abductions.

      "Speculate that the accusation is true, and that appropriate responses are applied."

      Under your scenario, applied by whom? The authorities? The church?

      "Those who argue that Wilson *and the parole board* are clearly in the wrong are in fact claiming that it would have been better that the child never be born than that it was put at risk of its father re-offending."

      That's like saying those who oppose polygamist cults are in fact claiming that it would have been better that the children never be born. Or that those who oppose rape are in fact claiming that it would be better that the child conceived in rape never be born.

      How we view an ethical decision at the stage of prevention, and how we view if after the chips fall, are two different considerations. By your logic, we can never say a human decision was clearly wrong inasmuch as it might always result some unintended good down the line.

      "A similar argument applies to vaccination. How does a 'high chance' of autism trade off against a 'low chance' of measles?"

      Only if you buy into the oft-discredited theory that vaccination causes autism.

      "Humans are notoriously bad at judging absolute risks and at comparing relative risks. What may be an obvious trade-off to one party may be a foolish trade-off to another."

      Which is one reason Christians need to have public debates about this issue. To hammer out thoughtful standing policies.

      Or are you just retreating into relativism? Simply because two people disagree on a risk assessment doesn't mean both positions are equally reasonable.

      "And if so, is it because the risks themselves are being mis-evaluated, or because priorities differ? There might be many layers of discussion involved to come to an agreement on highly contradictory 'obvious' positions."

      At the level of fact-free abstractions, naturally. But the actual issue is already far more specific: how should the church deal with the prospect of potentially dangerous parishioners? Should we take their word for it that they are not dangerous?

      Delete
    2. "Pedophiles pose a threat to any child who's a target of opportunity."

      So do unrecognised pedophiles. At what point do you want to shoot on sight?

      Yes, that's a reduction-ad-absurdum. It's also a completely logical consequence of saying "prevention at all costs". And if you're not saying "prevention at all costs", then you need a metric to decide how to balance a risk of recidivism versus genuine protection. Note that an unrecognised pedophile actually poses a far greater risk to your child than a recognised one; you just have much less information to evaluate the risk.

      Would I let a once-pedophile near my child? Sure. Would I leave them alone? Most likely not. Would I let them father a child of their own given a willing wife? Well, that's the tricky question, isn't it?

      "Only if you buy into the oft-discredited theory that vaccination causes autism."

      Except at this point I'm not discussing what *I* think. The point is that we need to both evaluate risks and compare risks, and both are really difficult in the general case. It doesn't help that some oft-discredited ideas turn out to be actually true, and others turn out to be foolish to entertain them in the first place. Given that (1) there have been documented instances of dangerous vaccines and (2) truly life threatening childhood infections like polio are outside the experiences of most people it's not hard to understand why people would feel more nervous about vaccination than infection, even if the best-practice mathematics suggest otherwise.

      The argument that known pedophiles pose a greater risk to children than unknown pedophiles falls prey to the same poor reasoning, in that it assumes that a comprehended risk is inherently more dangerous than an uncomprehended one.

      "Or that those who oppose rape are in fact claiming that it would be better that the child conceived in rape never be born."

      The offence in rape is against the woman, not the child. Where the child becomes an issue is when it is subsequently killed because it was conceived in the wrong way. (Unless one believes that the child inherits the evil of it's father, but that's a completely different discussion from where our society is at).

      Let me make one final remark on the case involving Wilson before I turn to the general case. A majority - but not all - of the supervisory board agreed that marriage was a positive move towards the man's rehabilitation. Assuming the subsequent crime, was that decision wrong, or was there a failure of supervision? If so, is this based on the benefit of hindsight, or was the evidence there already?

      More generally, there are two interesting questions arising:

      (1) Imagine that some number of pedophiles are rehabilitated post-incarceration and go on to be good fathers, and some do not. What statistical likelihood of re-offsense would one be willing to tolerate if the situation was generally good? If your answer is zero, how do you reconcile that with the knowledge that some proportion of (non-criminal) parents will end up abusing their offspring?

      Or does this have nothing to do with rehabilitation and is purely a matter of justice?

      (2) Given that there are many situations where some statistical proportion of outcomes will turn out well and some poorly, how much personal accountability accrues to those making the decisions, even though the decision itself is a moral choice of a 2nd party?

      As a safer example, if a recruiter recommends a person to a company and that employee ends up engaging in fraud some time later, how much fault accrues to the recruiter (assumption - he's only recommending one or two people, not a whole lot)? Does it make a difference if he was an ex-con, but all parties are aware of this going in?

      I think a certain amount of criticism should accrue to the recruiter for failing to predict, but that doesn't imply that the recruiter should take responsibility for enabling the crime.

      Delete
    3. Andrew W

      "Given that (1) there have been documented instances of dangerous vaccines"

      1. No reasonable person denies there are risks involved with vaccinations. Indeed, Steve explicitly states this in this very post.

      2. With this in mind, what "documented instances of dangerous vaccines" are you referring to?

      3. There have been "documented instances" that Andrew Wakefield's (an inspiration for your nom de plume?) supposed "research" linking MMR vaccinations with autism has been fraudulent. And Wakefield's work is quite arguably the predominant impetus and basis for the modern day anti-vaccination movement.

      "and (2) truly life threatening childhood infections like polio are outside the experiences of most people"

      1. Thanks in large part to vaccinations.

      2. Besides, the normal and primary recommendation today for polio vaccination isn't to be administered as a live vaccine. Rather, the normal and primary recommendation today is to administer it as a killed vaccine (i.e. the Salk inactivated poliovirus vaccine).

      3. You're wrong about a couple of facts:

      a. Generally speaking, polio isn't "life threatening." Although it can be in a minority of cases. What's usually most feared about polio is that it can lead to paralysis.

      b. Polio strikes adults just as well as kids. In fact, polio is usually worse in adults than in kids. For example, the chances of developing paralysis increases with age. Paraplegia, quadriplegia, and respiratory muscle dysfunction have been most prevalent in individuals over 15 years old.

      4. Polio has been all but eradicated in the US and indeed most of the Western world. At one point, many physicians and scientists thought polio could be eradicated around the world by 2005 (much like smallpox). But there are several pockets of resistance, especially in India, Pakistan, and Nigeria.

      However, more to the point, what do you think will eventually happen if, for example, today we immediately stop administering any and all polio vaccines around the world?

      "it's not hard to understand why people would feel more nervous about vaccination than infection, even if the best-practice mathematics suggest otherwise."

      Generally speaking, one would say people who continue to "feel more nervous about vaccination than infection" despite all the facts, evidence, and good reasoning are not thinking very rationally.

      Delete
  2. "How does a 'high chance' of autism trade off against a 'low chance' of measles?"

    Beg the question much?

    ReplyDelete