A “hazard” model is common statistical model used to explain rare events. In such a model, the chance of a certain kind of event happening (e.g. death) goes as the product of rate contributions from various factors. (E.g., a factor each for age, gender, income, smoking, etc.) Each value of each factor then contributes a “hazard ratio”, a factor by which that value increases or decreases the event chance, relative to some standard value.
There's also the issue that if someone secretly sympathizes with the offender (or is the offender), they will be more willing to blame the victim. For example, someone part of a frat-boy culture of sexual harassment may be more willing to blame the woman who was harassed, over the man harassing her.
I think we demand a higher level of caution and alertness from drivers in bad weather conditions. But we're also more lenient when something goes wrong. Or simply have difficulty telling who caused the wrong, when the conditions get sufficiently chaotic. I don't think, people often lose their drivers license, for causing accidents during a snow storm or on icy roads, even if they were found culpable. But I'd have to research that, maybe.
For a traffic accident we may blame the weather causing difficult driving conditions which may make B far more likely. For large values of N, we may be a bit more lenient with punishments for mid-levels of C and V.In traffic accidents, the V and C attributes often apply to varying levels to multiple people. We want to punish high levels of C and also encourage low levels of V. V is important, too. "Defensive driving" is a central concept in German traffic education, which you may call proactive victim blaming :)I think this works out reasonably well for us.We also make sure to discourage high levels of V outside of accidents by punishing people who just violate traffic laws (drunk driving, speeding, reckless behaviour etc.).
I would have thought that the simpler explanation is that we have a strong tendency to deny that things happen by chance and that outcomes are out of our control. We cling to the thought that we will be safe as long as we do x, y, and z. Hence, if a bad thing happens to a person, they must not have been taking whatever steps would have reliably prevented it.
I think it is acceptable to put some people in prison even though factors vary a lot, as long as the Criminal factor is drawn from a roughly-consistent-over-time distribution. Obviously it's more complex than that, because a moment to the next is not independent, but it's simple to explain this way.If C is drawn from a normal distribution, then a person we want to put in prison will have a distribution with at least one of two features:
(1) the distribution is high enough that it is frequently in the range that will lead to crime being committed if it coincides with other factors being high. (keeps an eye out for crimes of opportunity)
(2) the right tail is large enough that they occasionally have a period of unpredictable, ridiculously high, C. (prone to mental breakdowns, in a dangerous way).
One interesting point, distributions that come about from a product of many independent factors tend to have long tails, log normals. If the true distribution has many factors then a long tail will be present. To explain the rare, long tail events that do occur with some frequency within a simple two factor model might then require one to assume a large factor from one of the two components retained.
To blame is different from to explain. Simply because variation in X explains variation in B, does not mean that X is to blame for B.
In your traffic accident model, you give the example of N, nature, providing road conditions and bad weather. But it does us no good to blame nature, because it's not something we can change. We can't put it in jail, we can't threaten it with consequences if it doesn't change its behavior.
In your traffic accident model you also give the example of V, the victim. It often does us no good to blame V, either; perhaps V was following all the normal rules of traffic. Most victims, even completely law-abiding ones, could have prevented an accident if they were abnormally alert. We cannot demand that everyone be abnormally alert all the time. The rule would not succeed in making people abnormally alert. It could also cause harm, because if everyone is trying to be alert against crazy drivers coming from any direction against traffic rules, they may be less alert for the more typical things a driver should pay attention to.
We should blame someone for a crime only when it does us some good to do so. That means that by punishing the person, we actually, effectively reduce the undesired outcomes. And it means that our assignment of blame does not have excessive harmful side effects.
There's also the issue that if someone secretly sympathizes with the offender (or is the offender), they will be more willing to blame the victim. For example, someone part of a frat-boy culture of sexual harassment may be more willing to blame the woman who was harassed, over the man harassing her.
That makes a lot of sense too.
I think we demand a higher level of caution and alertness from drivers in bad weather conditions. But we're also more lenient when something goes wrong. Or simply have difficulty telling who caused the wrong, when the conditions get sufficiently chaotic. I don't think, people often lose their drivers license, for causing accidents during a snow storm or on icy roads, even if they were found culpable. But I'd have to research that, maybe.
For a traffic accident we may blame the weather causing difficult driving conditions which may make B far more likely. For large values of N, we may be a bit more lenient with punishments for mid-levels of C and V.In traffic accidents, the V and C attributes often apply to varying levels to multiple people. We want to punish high levels of C and also encourage low levels of V. V is important, too. "Defensive driving" is a central concept in German traffic education, which you may call proactive victim blaming :)I think this works out reasonably well for us.We also make sure to discourage high levels of V outside of accidents by punishing people who just violate traffic laws (drunk driving, speeding, reckless behaviour etc.).
I would have thought that the simpler explanation is that we have a strong tendency to deny that things happen by chance and that outcomes are out of our control. We cling to the thought that we will be safe as long as we do x, y, and z. Hence, if a bad thing happens to a person, they must not have been taking whatever steps would have reliably prevented it.
I think it is acceptable to put some people in prison even though factors vary a lot, as long as the Criminal factor is drawn from a roughly-consistent-over-time distribution. Obviously it's more complex than that, because a moment to the next is not independent, but it's simple to explain this way.If C is drawn from a normal distribution, then a person we want to put in prison will have a distribution with at least one of two features:
(1) the distribution is high enough that it is frequently in the range that will lead to crime being committed if it coincides with other factors being high. (keeps an eye out for crimes of opportunity)
(2) the right tail is large enough that they occasionally have a period of unpredictable, ridiculously high, C. (prone to mental breakdowns, in a dangerous way).
One interesting point, distributions that come about from a product of many independent factors tend to have long tails, log normals. If the true distribution has many factors then a long tail will be present. To explain the rare, long tail events that do occur with some frequency within a simple two factor model might then require one to assume a large factor from one of the two components retained.
To blame is different from to explain. Simply because variation in X explains variation in B, does not mean that X is to blame for B.
In your traffic accident model, you give the example of N, nature, providing road conditions and bad weather. But it does us no good to blame nature, because it's not something we can change. We can't put it in jail, we can't threaten it with consequences if it doesn't change its behavior.
In your traffic accident model you also give the example of V, the victim. It often does us no good to blame V, either; perhaps V was following all the normal rules of traffic. Most victims, even completely law-abiding ones, could have prevented an accident if they were abnormally alert. We cannot demand that everyone be abnormally alert all the time. The rule would not succeed in making people abnormally alert. It could also cause harm, because if everyone is trying to be alert against crazy drivers coming from any direction against traffic rules, they may be less alert for the more typical things a driver should pay attention to.
We should blame someone for a crime only when it does us some good to do so. That means that by punishing the person, we actually, effectively reduce the undesired outcomes. And it means that our assignment of blame does not have excessive harmful side effects.