Panera Bread makes non-statement re: guns in their stores

[xBiffx]

You most certainly wouldn't be proving a negative. If gun owners are safer due to their firearms they should have a lower rate of victimization from various crimes.

How can you prove that without proving how many times something didn't happen. Hence proving a negative. All you'd be proving by showing victimization rates is how many times an attempt was or was not stopped because of the presence of a firearm. You can't know how many times nothing happened because a would be attacker knew there would be a gun pointed at them or anticipated it so chose not to engage.

 

[fskimospy]

How can you prove that without proving how many times something didn't happen. Hence proving a negative. All you'd be proving by showing victimization rates is how many times an attempt was or was not stopped because of the presence of a firearm. You can't know how many times nothing happened because a would be attacker knew there would be a gun pointed at them or anticipated it so chose not to engage.

No, still not proving a negative. Let me break it down:

X% of Americans are victims of crime each year. If conduct a robust statistical analysis of crime trends among CCW holders that shows that 10% fewer CCW holders are victims of crime each year than the average American (controlling for demographics, etc), then you've shown a correlation between CCW holders and lower victimization rates.

 

[Zargon]

Well that certainly seems like a problem, doesn't it?

No? does it surprise you that people that don't own cars, get in less car accidents than those who do?

Is it weird that someone who doesn't own a chainsaw has never hurt themselves with one?

Has common sense completely left us now?

My point is simply that for the average person the research currently indicates that gun ownership makes you less safe. Maybe CCW holders are immune from this trend, but considering the magnitude of difference my guess is that's unlikely.

you are a big sucker for correlation == causation eh

in that case, hot weather causes crime, and semen makes women happy. yes, there are statistically significant resulted studies showing that women having unprotect sex thus having contact with semen are happier than those are not

and studies that correlate the temperature outside rising with a rise in crime rates

that doesn't mean that hot weather causes crime, or that semen makes women happy.


Not everyone is responsible, our highway fatality statistics prove it. Some of those people are going to own guns and do stupid things with it. that doesn't mean that everyone that owns one is now more likely to do something stupid.

There's a lot more to it than crime rates though such as accidents, suicides, etc.

suicides? you are implying that what, those people commited suicide PURELY because they own a firearm?

I mean, it probably had nothing to do with depression or other mental illness right? it was the eeeeeeeeeeeevil gun that made them do it?

 

[fskimospy]

No? does it surprise you that people that don't own cars, get in less car accidents than those who do?

Is it weird that someone who doesn't own a chainsaw has never hurt themselves with one?

Has common sense completely left us now?

Nope.

you are a big sucker for correlation == causation eh

in that case, hot weather causes crime, and semen makes women happy. yes, there are statistically significant resulted studies showing that women having unprotect sex thus having contact with semen are happier than those are not

and studies that correlate the temperature outside rising with a rise in crime rates

that doesn't mean that hot weather causes crime, or that semen makes women happy.

Not everyone is responsible, our highway fatality statistics prove it. Some of those people are going to own guns and do stupid things with it. that doesn't mean that everyone that owns one is now more likely to do something stupid.

I think you need to learn more about statistics and research design before talking about this. One of the fundamental elements of research is a plausible causal mechanism. I have no patience for people who say 'correlation doesn't mean causation' as a way to dismiss any research they don't like. Correlations mean a lot if they are part of a larger analytic framework.

suicides? you are implying that what, those people commited suicide PURELY because they own a firearm?

I mean, it probably had nothing to do with depression or other mental illness right? it was the eeeeeeeeeeeevil gun that made them do it?

No, just that guns are among the most effective ways to commit suicides. This is, for example, one of the reasons why women may have more suicide attempts than men, but men have more successful suicides. They use superior means.

 

[xBiffx]

No, still not proving a negative. Let me break it down:

X% of Americans are victims of crime each year. If conduct a robust statistical analysis of crime trends among CCW holders that shows that 10% fewer CCW holders are victims of crime each year than the average American (controlling for demographics, etc), then you've shown a correlation between CCW holders and lower victimization rates.

Victimization rates, that doesn't tell how many times they avoided victimization it tells you how may times they didn't avoid it. It just tells you on average, how often they may expect to be victimized. If the rate is 10% fewer that doesn't mean that they are avoiding 10% more victimizations. It could be a lot more than that, but you can't put a number on what didn't happen.

Another way to look at it, rates tell you how often to expect something and that doesn't mean how often you can expect the opposite of what the rate is telling you. Again, you can't put a number on something that never happened. That's proving a negative.

 

[fskimospy]

Victimization rates, that doesn't tell how many times they avoided victimization it tells you how may times they didn't avoid it. It just tells you on average, how often they may expect to be victimized. If the rate is 10% fewer that doesn't mean that they are avoiding 10% more victimizations. It could be a lot more than that, but you can't put a number on what didn't happen.

Another way to look at it, rates tell you how often to expect something and that doesn't mean how often you can expect the opposite of what the rate is telling you. Again, you can't put a number on something that never happened. That's proving a negative.

I'm sorry but this is all really wrong.

If you have a random sample of the population you are trying to measure then that sample should be normally distributed and therefore amenable to statistical analysis. To be clear there are certainly some controls necessary (like demographic dispersal, geographic dispersal, etc) that need to be accounted for as CCW holders are not evenly distributed throughout the country, but these can be accounted for.

After you do that, you look at victimization rates of the average American vs. the victimization rates of the average CCW holder. Or more clearly, look at the error term in your regression for the CCW variable and see if the p value is significant. That tells you if there's a negative correlation between having a weapon and being victimized less by crime. ie: telling you if being armed is preventing victimization.

 

[WackyDan]

Mine insures all my guns so they know how many and what type I own. Didn't change anything in my rates. You are again full of shit.

Also, risk is not equivalent to safety and risk is what the insurance company is looking at.

My rates didn't change one bit either.

 

[xBiffx]

I'm sorry but this is all really wrong.

If you have a random sample of the population you are trying to measure then that sample should be normally distributed and therefore amenable to statistical analysis. To be clear there are certainly some controls necessary (like demographic dispersal, geographic dispersal, etc) that need to be accounted for as CCW holders are not evenly distributed throughout the country, but these can be accounted for.

After you do that, you look at victimization rates of the average American vs. the victimization rates of the average CCW holder. Or more clearly, look at the error term in your regression for the CCW variable and see if the p value is significant. That tells you if there's a negative correlation between having a weapon and being victimized less by crime. ie: telling you if being armed is preventing victimization.

How many more times did a firearms owner avoid victimization vs. a none owner then? I'd be interested to know your answer considering the rates only tell you how many less times a firearms owner was victimized vs. the non owner. They don't tell you how many times they were not victimized. That number would be key in determining what owning a firearm did for them.

 

[Fayd]

Well Mr. internet troll... You entirely missed the point didn't you?

They either want to ban concealed carry in their stores or they don't. They should make a decision instead of playing the fence.

Why should they? they're a business. It's their decision what kind of official message they want to send (or not send).

unless you're a stockholder, you have no part to play in that decision.

 

[Paratus]

You are silly. You just defeated your own argument. If no guns allowed signage is not going to put me in harms way because the odds of someone else killing me is so low, than why ban me carrying in the first place? By your own words, the chance of one of these incidents happening is low. The odds of me needing a gun to defend myself, not just from a mass shooter but from regular mugging/robbery/etc is far higher than the odds of a legal CCW holder doing something illegal (which is pretty much unheard of).

Ever hear of a CCW holder accidentally discharging their weapon in public? Is it maybe possible that the president of Panera feels the risk of an accidental discharge at a Panera is greater than the benefit of an armed patron possibly stopping a robbery?

 

[fskimospy]

How many more times did a firearms owner avoid victimization vs. a none owner then? I'd be interested to know your answer considering the rates only tell you how many less times a firearms owner was victimized vs. the non owner. They don't tell you how many times they were not victimized. That number would be key in determining what owning a firearm did for them.

I'm not sure what else to say. Because in our theoretical analysis we have a normally distributed population, a random sample, and we have controlled for other possible sources of variance that's exactly what it will tell them.

 

[Zargon]

My rates didn't change one bit either.

mine might, but only because eventually you have to add extra coverage if you end up with too many

 

[xBiffx]

I'm not sure what else to say. Because in our theoretical analysis we have a normally distributed population, a random sample, and we have controlled for other possible sources of variance that's exactly what it will tell them.

Nope, sorry, you've got no data on how many times a firearm owner wasn't victimized, only data on how many times they were. So drawing a conclusion about the roll the firearm played, specifically how it can aid in avoiding victimization, is haphazard.

 

[fskimospy]

Nope, sorry, you've got no data on how many times a firearm owner wasn't victimized, only data on how many times they were. So drawing a conclusion about the roll the firearm played, specifically how it can aid in avoiding victimization, is haphazard.

Like I already said, random sampling and a normally distributed population cover that. In a random sample of a normally distributed population all people with similar characteristics except for a CCW will be victimized by crime to an extent that is distributed where ~95% of people fall within two standard deviations of the mean. If when you analyze those with CCW permits and their victimization falls more than two standard deviations above or below the mean you can say with high confidence that it is in fact aiding in avoiding (or possibly causing more!) victimization.

You aren't arguing against my position, you're arguing against math.

 

[Rebel_L]

Why should they? they're a business. It's their decision what kind of official message they want to send (or not send).

unless you're a stockholder, you have no part to play in that decision.

They should because they are being very unclear. The company has come out and says that they don't want guns in the stores, but at the same time they say they are not going to let their customer know about their wishes as they will not be posting signs in the stores or having employees talk to customers about it. Its a very mixed message, say one thing, do another, asking them to clarify their position by having some sort of consistent behavior would seem like a reasonable action for anyone who is considering being a customer.

 

[xBiffx]

Like I already said, random sampling and a normally distributed population cover that. In a random sample of a normally distributed population all people with similar characteristics except for a CCW will be victimized by crime to an extent that is distributed where ~95% of people fall within two standard deviations of the mean. If when you analyze those with CCW permits and their victimization falls more than two standard deviations above or below the mean you can say with high confidence that it is in fact aiding in avoiding (or possibly causing more!) victimization.

You aren't arguing against my position, you're arguing against math.

You are ignoring what the numbers tell you, and assuming what they don't tell you. The numbers tell you victimization rates, not non victimization rates. The non rate would be key in knowing what role a firearm plays. Just because they may have been victimized say 10 times whereas non owners were victimized 20 times that doesn't tell you how many time the firearms owners weren't victimized. All it says is that they weren't victimized 10 times that the non owners were. For all we know, they could have avoided 2, 20, 200 more victimization. There is no way to know because the victimization didn't happen. All we can know is the frequency they were victimized and not when they weren't.

That's not arguing against math, one fucking bit.

 

[fskimospy]

You are ignoring what the numbers tell you, and assuming what they don't tell you. The numbers tell you victimization rates, not non victimization rates. The non rate would be key in knowing what role a firearm plays. Just because they may have been victimized say 10 times whereas non owners were victimized 20 times that doesn't tell you how many time the firearms owners weren't victimized. All it says is that they weren't victimized 10 times that the non owners were. For all we know, they could have avoided 2, 20, 200 more victimization. There is no way to know because the victimization didn't happen. All we can know is the frequency they were victimized and not when they weren't.

That's not arguing against math, one fucking bit.

Normal distribution.

Random sample.

You are arguing against math.

 

[xBiffx]

Normal distribution.

Random sample.

You are arguing against math.

Nope, I'm not. I'm arguing what the numbers are telling you, you are arguing what they don't tell you. That's not math.

Easy question, how many times was a firearms owner not victimized? Should be easy, its just math, amight?

 

[Rebel_L]

You are ignoring what the numbers tell you, and assuming what they don't tell you.

That's not arguing against math, one fucking bit.

The numbers tell you how often people with guns get victimized and how often people without guns get victimized. The difference between the two is a measure of how much "less" you are victimized with a gun. I really don't see what is so hard to understand about that, victimization is not something you avoid, its something that happens. Getting mugged is an event that happens, not getting mugged is not an event.

If you wanted to argue for the benefits of gun ownership in the name of safety you really should look at the unquantifiable benefits, like added peace of mind. It may be irrational to feel safer having a gun, but that does not change that many people do, and this can make it a rational decision to have a gun as those benefits can outweigh the added risk of having a gun around.

 

[xBiffx]

The numbers tell you how often people with guns get victimized and how often people without guns get victimized. The difference between the two is a measure of how much "less" you are victimized with a gun. I really don't see what is so hard to understand about that, victimization is not something you avoid, its something that happens. Getting mugged is an event that happens, not getting mugged is not an event.

If you wanted to argue for the benefits of gun ownership in the name of safety you really should look at the unquantifiable benefits, like added peace of mind. It may be irrational to feel safer having a gun, but that does not change that many people do, and this can make it a rational decision to have a gun as those benefits can outweigh the added risk of having a gun around.

The difference isn't how much less. A gun owner avoids victimization, where is that recorded? Just subtracting owners rates from non owners rates does not for a second tell you how many times the owner wasn't victimized. It only tell you how many times they were victimized compared to how many times a non owner was victimized. For all we know, they avoided 10 times that many but you can't know it because you can't show in numbers something that never took place. Hence attempting to prove a negative.

I'd agree, looking at numbers is only half the reasoning, but eskimo seems to want to get stuck on arguing that the numbers are telling him something that they in fact aren't. There is no rate for how many times a gun owner ins't victimized. That's the same as saying there is not rate for how many times something doesn't happen. That's a simple concept to grasp.

 

[fskimospy]

Nope, I'm not. I'm arguing what the numbers are telling you, you are arguing what they don't tell you. That's not math.

Easy question, how many times was a firearms owner not victimized? Should be easy, its just math, amight?

Rebel_L is also trying to explain this to you. You are arguing against math.

Let's look at some made up numbers for ease of explanation: (and we will assume guns lead to less victimization)

Since our population is normally distributed and our sample is random, assuming a sufficient sample size the average person in our sample and the average person in the population will have an approximately equal number of victimizations.

According to our analysis the average american will be victimized by crime 10 times in a year with a standard deviation of 1 crime. If, after accounting for other salient variables, a CCW holder is victimized 7 or fewer times in a year (ie: more than 2 standard deviations from the mean) then we can say that gun ownership is highly likely (in this case ~97% likely) to lead to a lower rate of victimization. In this case we could also draw up a confidence interval to show you where the mean number of victimizations for a CCW holder should fall.

The number of avoided victimizations would be the difference between that mean and the population mean.

Do you understand now?

 

[xBiffx]

The number of avoided victimizations would be the difference between that mean and the population mean.

That doesn't tell you how many times they avoided victimization. It tells you how many less times they were victimized. They aren't the same thing. Once could be much, much more than the other but there is no way to know.

The same goes for the average population. You have a number for how much something happens. Not a number for how much it doesn't. So when you compare two numbers for different groups that tell you how much something happens, it can never tell you how much something doesn't happen. That is simple.

If you have an average of 10 victimizations per year, how does that tell you how many times they weren't victimized.

 

[fskimospy]

That doesn't tell you how many times they avoided victimization. It tells you how many less times they were victimized. They aren't the same thing. Once could be much, much more than the other but there is no way to know.

The same goes for the average population. You have a number for how much something happens. Not a number for how much it doesn't. So when you compare two numbers for different groups that tell you how much something happens, it can never tell you how much something doesn't happen. That is simple.

This is incorrect.

Variations like you describe are accounted for by the normal distribution and random sampling.

 

[Rebel_L]

The difference isn't how much less. A gun owner avoids victimization, where is that recorded? Just subtracting owners rates from non owners rates does not for a second tell you how many times the owner wasn't victimized. It only tell you how many times they were victimized compared to how many times a non owner was victimized. For all we know, they avoided 10 times that many but you can't know it because you can't show in numbers something that never took place. Hence attempting to prove a negative.

You cant avoid becoming a victim, you can only become one. That aside what is important about the number of times you "avoided" being a victim other than in comparison to something else? If you avoided 99 times becoming a victim in a day because you had a gun but were mugged once anyways, and I avoided 98 without one and got mugged twice the only thing that dealing with the other incidences of non events can do is make your position weaker. You had 2% avoidance rate and I had a 1% avoidance rate, that only gives you a 1% difference. If we take out all the missed muggings you are all of a sudden 50% less likely to be victimized than me.

 

[xBiffx]

This is incorrect.

Variations like you describe are accounted for by the normal distribution and random sampling.

No they aren't. They aren't included in the numbers at all. The numbers are telling you something completely different.

It's not a variation. It the opposite of what you are recording. Number of victimization vs. number of avoided victimizations.