# Individual actions vs. statistical likelyhood

Obviously inspired by, but not necessarily limited to, the gun debate. My question here is, when should human behavior be regarded as the actions of individuals, and when should it be treated statistically as the population as a whole? To use the gun debate example, here’s a prototypical dialogue:

Person A:“A gun in the home is twenty times more likely to kill its owner or a family member than an intruder”
Person B: “I’m not mentally ill, a criminal or a moron. I’m not going to commit suicide or murder my family, and I religiously follow safe handling and use procedures”.
Person A: “Well, maybe you do. But what about all the people who don’t?”
Person B: “Why punish me for what others do wrong? Whatever happened to personal responsibility?”
And so on…

In general, conservatives are more likely to regard individual behavior and personal moral responsibility as the basis of society, and liberals are more likely to see things in terms of impersonal forces such as one’s social environment acting on groups of people as whole. Human behavior tends to straddle the chaotic line between mechanistic explanations and statistical explanations. So where is it appropriate to draw the line?

Statistics are useful for both summarizing and for predicting.

Predicting stuff is much more difficult.

Predicting what a given individual will do is like predicting what a given stock price will be at noon, 1.6 years from now. Sure, it’s possible to guess, but the assumptions that go into the prediction equation tend toward the subjective when there’s a lack of data (because it’s unmeasurable or poorly understood).

I’d suggest that statistics, as a discipline, has general guidelines about drawing that line, but not specific ones. Overfitting a predictive model makes for poor predictions outside of the model’s input parameters and the data on which it’s based. Underfitting makes a model perform poorly because it is too simplistic.

Even reasonable, objective statisticians may disagree on what the “best” model is and where to draw that line sometimes, but not due to lack of effort or research into model-building.

Your premise is correct in my opinion: if someone claims that people are deterministic, like “put them in Environment A, add a pinch of Economic Situation B, stir in Social Factor C, and expose them to Crisis D, then you get Y”, that will likely be a false statement in some cases.

However, there is some utility when interventions may improve things based on measurable endpoints. How may one compare? Usually history is a good benchmark. Or if there is an increased and durable association with positive, desirable things.

Statistical prediction performs well under nice, controlled conditions. It’s difficult to draw those lines on a given social issue because human societies are one vast uncontrolled experiment.

I think this really is the crux of what it means to govern. It’s always going to be a balance between the wellfare of the public as a whole, determined statistically, and the rights of the public as a set of individuals.

It’s like the old moral quandry of killing one to save a million. Oversimplifying perhaps, while no one would like to make that decision, I think if forced to do so, virtually everyone would choose to kill the one. But that’s an easy example of the statistics over the individual. That decision gets a lot harder when that number gets larger. What’s the maximum number of people you would kill to save a million? 5? 500? 500,000?

That’s what we face in a much more obscure way when we talk about something like gun control. We have statistics about the dangers and benefits of guns. Even assuming that both sides could agree on what those statistics say about the dangers and benefits, we’re still stuck weighing that against the definite costs of the rights of the individuals. Some people will argue that even one death from guns isn’t worth those rights of the millions of gun owners, and there’s plenty of views in between. And, of course, we can’t even agree on what the statistics say about how many lives guns cost or save or whatever.

So, sure, it’s easy when it’s something like seat belt laws, where the statistics are pretty clear that it saves lives and the cost is the right to not wear a seatbelt; you’re not going to get a lot of pushback on that since not many people hold that right as particularly important. But in most cases, drawing that line is a lot more difficult. But, in general, you’re going to see various people put the line in different places depending on how they feel about individual freedom vs. general wellfare as well as their personal value on the topics involved.

And as it comes to the gun debate, I think it’s even more difficult than that, because it’s also a highly emotionally charged event. Attempting to be as non-biased as I can, whatever the costs and benefits are, in general, most pro-gun people felt that line is on the side of gun rights, and most anti-gun people felt those costs put it on the other side. So when we have a highly public incident like Newtown, I think it galvanizes people and they lose the perspective that, as tragic as those deaths are, there are thousands of other deaths related to guns every year, and talk from either gun control people to have more restrictive laws or pro gun people wanting less restrictive laws, both claiming the goal of preventing that sort of event, are really losing perspective on the issue as a whole.

If gun owners were only killing themselves, there wouldn’t be a debate. It’s fine to argue about your own individual liberty; the problem comes from depriving others of theirs.

The stats on the elevated risk for death for those in the home due to having a gun, relative to needing to fend off an.attack is why I don’t have a firearm. The stats on numbers of other people killed by firearms is why I think there.should be restrictions on other people’s ownership of firearms.

People who advocate for unrestricted firearm ownership are in part responsible for the unsettled feeling the rest of us now have sitting down in a movie theater or sending our kids off to school.

Your individual liberties don’t trump everyone else’s.

No, your own ignorance of statistics is responsible for that. Unless you’re similarly unsettled because your kids might get struck by lightning or contract the plague, both quite a bit more likely than dying in a mass murder. And I bet you don’t even give them lightning rod helmets and face masks when you send them to school, do you? Do you advocate for such things politically? Both would save far more lives than any type of gun control.

Don’t be fooled into thinking “fear” is the same thing as “statistics”. Fear of Aurora/Sandy Hook style massacres is simply not rational, and should absolutely not be the basis behind any legislation.

If you’re going to use statistics, use them. Don’t just pull them out when you’re scared and your pet bugaboo needs to be scolded again.

I can assure you that I am not particularly ignorant of statistics. In fact, in all likelihood I understand them quite a bit better than you, given my training and experience and daily use of statistical techniques.

Tell me how many people are shot on average in the US each year, versus being struck by lightning or contracting the plague. For someone who wants to throw down regarding statistics, you are painfully, woefully ignorant of some basics here.

The fact is, even if you just restrict it to people in the US who are killed in mass shootings versus people who die from lightning strike, you’re going to be embarrassingly wrong. The US averages about 20 incidents of mass shootings per year versus about 40 fatalities due to lightning strikes. 68 people died in mass shootings in the US in 2012. 68 is more than 40. Approximately 100,000 - 110,000 people in the US will be shot in a given year. This number is also greater than 40.

It’s comical to be lectured by a gun advocate about not being motivated by fear.

Please see above, and note that in your effort to employ statistics yourself, you a) provided no actual statistics, and b) were simply wrong.

This is exactly why making policy around these sensationalist stories, either pro or anti gun, is ridiculous. Taking your number of 68 as accurate, I’m about 52x more likely to die to accidental drowning (cite: CDC). How often do we actually have national discussions about water safety? Do you fear for your life when you take a bath? Should we pass laws to make bathing safer to reduce those deaths? Or, contrary, do we have groups fighting to increase our water rights against oppressive safety laws?

And that’s just a silly example. What about real issues facing this country, like deaths due to obesity, pollution, cancer? We’re talking about changing laws over 68 deaths in a year and not taking something that causes hundreds of thousands of deaths as seriously at a national level. Why is the line SO much different for this issue compared to others?

And really, why should we live in fear of something like that? I live in the DC area and I remember when we had the sniper. I remember seeing tons of people crouching behind their cars when pumping gas, rushing from their cars to the stores to be outside as little as possible. I remember when a witness lied that they saw him shoot out of a white van and then people being scared whenever they saw one, even though they’re pretty much everywhere. There’s millions of people in the area where that happened, and there were 14 total victims. So even when it was going on, I didn’t change my behavior at all. I pumped gas like usual, I wasn’t scared of vans. The chances were just too low to be worried about. But yet people were flipping out over it.

My point is, based on the question in the OP, if we’re going to make governing decisions based on statistics of populations, talking about 68 people out of 315M, or about 0.02 per 100k, we’re talking about a drop in the bucket compared to the total number of deaths (cite: CDC). So if that’s the place where the line should be drawn, we should be reconsidering a lot of laws in a lot of other areas too. Otherwise, it seems to me that it’s an outlier.

Ugh. Won’t someone think of the tortured analogies? Why do people constantly need to point to some more common occurrence and say that because of it, we shouldn’t do anything about gun violence? Trust me, you are not the first person to get exercised about drownings. If it’s a meaningful and relevant comparison, I suppose we should ensure that there is a lifeguard watching over people while they use firearms.

[I shouldn’t pursue this fruther, because every discussion is derailed by these fucking tiresome analogies, but they fail on several basic levels: 1) there is generally no opposition to taking steps to reducing the risk of harm for essentially any other risk factor (e.g. there are laws and recommendations to improve bath safety), 2) most of the commonly cited conditions cannot be foisted upon someone else (e.g. I am not at risk from someone putting me in a bathtub while I am sitting at the movie theater) or can be avoided by people who estimate the risk to outweigh the benefit, and 3) the rate of poor outcomes on a per use basis inevitably dwarfs the poor outcomes associated with firearms (e.g. most everyone in the whole population uses a bathtub on a daily basis, whereas only a minority use a firearm on a daily basis).]

In terms of the numbers, I’m sure that the 100,000 people who are shot per year in the US would rather not be shot. Is 100,000 a meaningful number to you?

The reason why spree killings have more impact than their raw numbers is because we are human beings. We are motivated to act by the degree to which we can put ourselves mentally in a particular circumstance. Case by case, most people in the US can identify how a given gun death happened in circumstances unlike our own. We can fantasize ourselves out of those conditions in order to reduce the risk. On the other hand, we cannot imagine ourselves not going to school or not sending our kids off to school, and can imagine ourselves regularly going to a movie theater or to a mall or to work or to a small rally for a local political candidate or to a dance or to court or to workout at the gym or to McDonald’s and so on and so forth.

In these circumstances, most people can identify with the victims, and most people cannot identify with people who want to have an AR-15 or AK-47. So, the equation is not in your favor, even if the numbers might otherwise be hand-waved away.

The answer to person A’s final question is that they go around claiming that “I’m not mentally ill, a criminal or a moron. I’m not going to commit suicide or murder my family, and I religiously follow safe handling and use procedures”.

This is true even outside the gun debate. Almost everyone on the road believes that they are a good driver, yet the sheer number of idiots on the road indicate that they can’t all be right. It is human nature to believe the unlike everyone else you really can drive home drunk just fine. Those safety regulations aren’t necessary for you, and that it is really everyone else that is the problem. Unfortunately statistics are more likely to be right than an individuals judgement.

Human behavior is not ergodic. If I never get into a car, then the odds of my getting into an auto accident are not the same as the odds computed from the general public. There is no “natural frequency” of car accidents. Most events that we really care about are a combination of natural frequency and “endogeneity”, that is, how actions we take affect the outcome. There is a natural frequency of death by heart attack, but individual behaviors can be risk factors that can swing the odds a bit. But if a person dies of a heart attack, it may be impossible to tell whether a risk factor or just dumb luck caused the death in any meaningful sense.

I don’t think there is some objective, logical place to draw the line on an issue like guns. People who are deeply concerned about their proliferation just have to convince gun owners to give them up or accept greater restrictions. The statistics might be very accurate about describing the state of the world, but they do not come ready made with a policy interpretation.

You’re completely missing the point I’m trying to make. The OP is specifically about statistics vs. individuals. I’m not bringing up drowning vs. guns as a tired old argument, which it is, because I’m pointlessly reiterating it. I’m specifically trying to make the point that if we’re going to make government decisions based on statistics and individual rights, we shouldn’t be changing the standard. A thousand preventable deaths are a thousand preventable deaths, whether they were from guns, drunk driving, texting while driving, gluttonous consumption, or whatever. From a statistical perspective, the only thing that should matter is that they’re dying, that we’re accurate about the cause, and that we’re confident about how those laws will prevent some amount of those deaths. If we’re going to govern effectively by statistics, we need to divorce the effect of the cause of those deaths from the numbers because, inherently, there’s going to be baggage associated with guns that is different from alcohol, cell phones, cars, and cheeseburgers.

You’re pursuing a strawman and hijacking a thread into a gun debate. I don’t have any particular interest in a gun debate because there’s other threads for that and I’ve put my thoughts there.

Also, since this is a thread specifically about statistics, I’m sure you just forgot to provide dependable cites, particularly for your claim in point 3. I particularly cannot agree without a citation that only a minority are exposed to guns on a daily basis. It seems intuitively obvious to me that a huge number handle guns directly, and as a result, an even larger number are exposed indirectly. For instance, one household has a gun, everyone who lives in that house or is a neighbor is potentially affected by that gun owner’s safe use of that gun. Similarly, someone is carrying, concealed or not, everyone they are in proximity to throughout the day is potentially affected by that person’s safe use of that gun.

Speaking for myself, I live in Northern Virginia, where open carry is completely legal and concealed carry requires a permit, I’m also often around police, military, security guards, and other carrying citizens just through a normal day. So, despite not owning a gun myself, I’m exposed to the potential dangers of guns virtually daily, often many times a day. I’m sure this is different for people in different areas, but I think my point remains that you’re vastly underestimating how much people are exposed to guns, by saying a minority.

I think that’s a fair assumption that people that get shot don’t want to get shot. I don’t see what relevance that has to anything. BTW, you’re way off on your number (cite: http://www.cdc.gov/nchs/fastats/injury.htm).

It sounds like you’re arguing that we should deliberately make irrational laws because humans are irrational. Yes, the public overreacts to events, part of that is human nature and part of that is sensationalism by the news looking for ratings. However, I think it is the obligation of government to not be swayed by anything but the actual numbers when weighing them against the rights of individuals. I shouldn’t have my rights taken away because people irrationally overreacted. And, by the same token, I shouldn’t be underprotected by the government because people underreacted to the available statistics.

And this just illustrates my point. My ability to empathize or not with a victim shouldn’t be a reason to ignore the numbers. For instance, I cannot personally empathize with the obesity problem in the US because I work very hard to maintain my health, but looking at the numbers, I realize it’s a very real problem that needs to be addressed. By the same token, I CAN sympathize with people who are afraid of being shot in a theater or at school or at work, but looking at the numbers, I realize it’s an irrational fear and there’s probably other issues that will have a larger affect on more people.

I tend to think smoking is a more apt analogy, though still not great. Smoking imparts a externality towards others in a similar way that gun control advocates claim gun ownership does. The biggest difference is the damage inflicted is generally not immediate except maybe in the case of fires. Still - smoking and smoking related costs to individuals and society as a whole are significant and any efficacy gained from smoking through enjoyment or other means can be argued to be outweighed by the societal costs. And there is no constitutional barrier to banning smoking nationwide.

Apt analogy?

Sounds like the Brady Campaign definition of “mass shooting” instead of a more traditional one that most layman would recognize.

You can do a lot to mitigate your risks of getting shot too. To start with, you could take steps to avoid young black males, drugs, and Chicago. Those alone would see your risk of getting shot plummet dramatically.

ETA: I know some of you are thinking, “but Chicago’s basically banned guns already, shouldn’t it be safe?” and the answer is that no, unfortunately, the liberal dream for a gun-free utopia hasn’t panned out like we’d hoped.

And 213 people were struck by lightning last year. Your kids have their lightning rod helmets on, right?

And only 323 of those people were killed by a rifle. (That’s about 1 in a million, equal to the odds of being struck by lightning in a given year, on average.) So what is this assault weapons ban, high capacity magazine business all about?

Gun control doesn’t work. But if it did, the people targeting handguns are on the right track. Worrying about big scary guns because of a few high profile massacres is just as foolish as I originally said. Basing policy on Newtown and Aurora is foolish and ignorant.

I don’t own a gun.

I should say you’re right that I didn’t bring statistics in my first post. Also, I was wrong about the plague, but given the frequency, if I lived out west, I’d be about equally afraid of the plague as mass murders.

We have a lot of water safety education, in schools, at the beaches, at swimming pools, etc. Paramedics give lectures to schoolkids. There are PSAs on tv urging swimming and boating safety.

I’m a bit long in the tooth, and, yes, I do worry about my safety when in the bathroom. I watch my step, and have laid down non-slip safety surfaces.

There are TONS of laws regarding bathroom safety; talk to any house-builder about it.

Another staggeringly ignorant argument. Chicago, DC, and NYC are all localities with strong gun bans…all geographically very close to other localities with very lax gun laws. People in DC just drive a few miles to Virginia and then illegally transport guns home again.

This, by the way, also puts the lie to the NRA’s claim that gun laws “only affect law-abiding citizens.” No, they would also affect the thousands of law-breaking citizens who buy guns legally in one state and transport them to another.

(And the battle of the simplistic talking points marches steadily onward. P-K4.)

You seem to be focused on A statistic, namely the frequency of an event. The OP was not, and included inferential statistics as well. When we make decisions about what to do, what regulations to establish, we look at much more than just the raw frequency.

Gun advocates have done everything they can to limit the availability of data. However, recent polling evidence shows that only a minority of American people own a firearm. I feel quite safe in assuming that a majority of the American public uses some kind of bathing fixture every day.

The issue is private firearm ownership. You’re just reaching here.

You seem to be linking to a CDC site with a mortality statistic
If you want the number of people who are shot in a given year from CDCs data, you need to go to WISQARS and look for nonfatal injury data as well.

Wanting to reduce the number of people who are shot is not irrational.

You do know that 100,000 is greater than 213, right?

Uh, lighting rods do not reduce the risk of lightning striking that spot. They increase it.

The word “advocate” has a different meaning than the word “owner.”

You do know that 100,000 is greater than “between 1 and 7” right?

What does any of that have to do with the current uproar over assault weapons and school shootings? I even said that while I don’t agree with gun control, focusing on handguns is the statistically honest way to go. People don’t kill each other with rifles any more than they get struck by lightning.