Current events, goings-on in Delaware and anything else of interest here.
 #110152  by David
 
Image

Article: http://www.businessinsider.com/gun-crim ... ate-2017-6

Conclusion: It’s difficult to separate out the effectiveness of gun laws from the inherent stability found in a wealthy, educated society versus the instability that is endemic to its opposite. In this sense, the disparity in the rate of gun deaths is reflective of the widening gap between the haves and the have-nots in America. With gun violence, as with so many other societal ills, residents in suburban Massachusetts and rural Louisiana may as well live in two different countries.
 #110155  by California_Exile
 
That study is irresponsible statistical garbage from the gun ban people. Here's why:

Number one, separating murders committed with guns from murders committed by other means is inane, as we all know. And pretending that "number of firearm laws" -- regardless of what those laws say -- is a meaningful cause of anything makes zero sense, even if they can cook up some kind of correlation.

Number two, it ought to be a big red flag that the chart showing change in homicide rate excludes four low-population states because the rates were too small to report (that's Hawaii, Maine, North Dakota and South Dakota), and then the next three eye-popping changes are all low-population states too (Delaware, Alaska and Nebraska). And Rhode Island shows a huge change in the other direction. So 8 of the 13 lowest-population states are either excluded because the data aren't meaningful or show wild swings. Obvious sample bias problem, but why should that get in the way of a good headline?

Number three, and most importantly, the numbers don't make sense. The gun banners' report says the "gun homicide" rate in Delaware in 1991 was 2.15 per 100,000. Population in 1991 (per the FBI's Uniform Crime Report) was about 680,000, so that's 14 or 15 "gun homicides" in Delaware in 1991. But the FBI UCR also says there were 37 murders or non-negligent homicides (by any means) in Delaware in 1991, for a murder rate of about 5.8 per 100,000 population (up from 33 and 5.4 per 100,000 in 1990, BTW; small sample bias again). So, implicitly, the "2.15 gun homicides per 100,000 population in 1991" number means that there were 22 or 23 murders or non-negligent homicides in 1991 committed by means other than firearms. That really doesn't sound right, even if (back to point 1) you think there's a meaningful difference between murders-by-firearm and murders-by-any-means.

Fast forward to 2015. The FBI's UCR app doesn't include 2015 data yet, but you can see the table here:

https://ucr.fbi.gov/crime-in-the-u.s/20 ... s/table-20

In 2015 in Delaware, 63 murders, 52 committed with guns. (Gun banners have 53 gun homicides, but that's at least pretty close.) Gun banners say the population was about 946,000, which they say is a rate of gun homicides per 100,000 population of 6.12. Except 53 divided by 9.46 is about 5.6, not 6.12, so their math is wrong. The overall murder rate (including non-firearm murders) is 63 divided by 9.46, or 6.7 per 100,000 in 2015.

By comparison to 2014 (last year of stats in FBI UCR app), 54 murders in a population of 935,000, or 5.8 murders per 100,000 population. Underlying table shows 39 out of those 54 murders in 2014 were committed with guns, which would make the "gun homicides per 100,000 population" rate about 4.2.

https://ucr.fbi.gov/crime-in-the-u.s/20 ... s/table-20

So in conclusion: (1) Delaware had a spike in murders in 2015 relative to 2014, 9 more total murders, 13 more murders committed with firearms. An outsized proportion of that can be attributed to Wilmington (23 murders in 2014 to 26 in 2015, per the News-Journal); small sample bias, but a third of the statewide increase attributed to a city with 8% of the statewide population. I can't find stats for the "greater Wilmington area," but I can guess.

(2) That 2.15 gun homicides per 100,000 population number from 1991 can't be right, unless long-time residents remember 1991 as the Year of the Axe-Murderer or something. The overall homicide rate was 5.8 per 100,000 in 1991, still (or more accurately, again) 5.8 per 100,000 in 2014, and 6.7 per 100,000 in 2015. None of these numbers cover Delaware in glory, but the overall murder rate bounced up and down between 1991 and 2014 but wound up the same. Then it spiked up by 15.5% between 2014 and 2015 (or between 1991 and 2015, if you prefer to look at it that way, but again, small sample size and lots of fluctuation from year to year).

(3) Hard to overemphasize just how subject to small sample bias the homicide rate in Delaware is. Look at the FBI's UCR table, which goes back to 1960, and you'll see the rate has been as low as 2.3 per 100,000 in 2003 and as high as 10.3 per 100,000 in 1974. It bounces around a lot from year to year.

(4) While relatively little of the statewide change in murder rate is the City of Wilmington proper, the murder rate there is still in the 30s per 100,000 residents. That's South Africa and Guatemala territory.

(5) And most importantly, gun banners can't do math, are sloppy and don't mind lying to push an agenda. I wouldn't use this study for toilet paper.
 #110159  by Boots
 
Since the oversize image blew out the right margin, here is a more readable version:
That study is irresponsible statistical garbage from the gun ban people. Here's why:

Number one, separating murders committed with guns from murders committed by other means
is inane, as we all know. And pretending that "number of firearm laws" -- regardless of what
those laws say -- is a meaningful cause of anything makes zero sense, even if they can cook up
some kind of correlation.

Number two, it ought to be a big red flag that the chart showing change in homicide rate excludes
four low-population states because the rates were too small to report (that's Hawaii, Maine, North
Dakota and South Dakota), and then the next three eye-popping changes are all low-population
states too (Delaware, Alaska and Nebraska). And Rhode Island shows a huge change in the other
direction. So 8 of the 13 lowest-population states are either excluded because the data aren't
meaningful or show wild swings. Obvious sample bias problem, but why should that get in the
way of a good headline?

Number three, and most importantly, the numbers don't make sense. The gun banners' report
says the "gun homicide" rate in Delaware in 1991 was 2.15 per 100,000. Population in 1991 (per
the FBI's Uniform Crime Report) was about 680,000, so that's 14 or 15 "gun homicides" in
Delaware in 1991. But the FBI UCR also says there were 37 murders or non-negligent homicides
(by any means) in Delaware in 1991, for a murder rate of about 5.8 per 100,000 population (up
from 33 and 5.4 per 100,000 in 1990, BTW; small sample bias again). So, implicitly, the "2.15 gun
homicides per 100,000 population in 1991" number means that there were 22 or 23 murders or
non-negligent homicides in 1991 committed by means other than firearms. That really doesn't
sound right, even if (back to point 1) you think there's a meaningful difference between murders-
by-firearm and murders-by-any-means.

Fast forward to 2015. The FBI's UCR app doesn't include 2015 data yet, but you can see the
table here:

https://ucr.fbi.gov/crime-in-the-u.s/20 ... s/table-20

In 2015 in Delaware, 63 murders, 52 committed with guns. (Gun banners have 53 gun homicides,
but that's at least pretty close.) Gun banners say the population was about 946,000, which they
say is a rate of gun homicides per 100,000 population of 6.12. Except 53 divided by 9.46 is about
5.6, not 6.12, so their math is wrong. The overall murder rate (including non-firearm murders) is
63 divided by 9.46, or 6.7 per 100,000 in 2015.

By comparison to 2014 (last year of stats in FBI UCR app), 54 murders in a population of
935,000, or 5.8 murders per 100,000 population. Underlying table shows 39 out of those 54
murders in 2014 were committed with guns, which would make the "gun homicides per 100,000
population" rate about 4.2.

https://ucr.fbi.gov/crime-in-the-u.s/20 ... s/table-20

So in conclusion: (1) Delaware had a spike in murders in 2015 relative to 2014, 9 more total
murders, 13 more murders committed with firearms. An outsized proportion of that can be
attributed to Wilmington (23 murders in 2014 to 26 in 2015, per the News-Journal); small sample
bias, but a third of the statewide increase attributed to a city with 8% of the statewide
population. I can't find stats for the "greater Wilmington area," but I can guess.

(2) That 2.15 gun homicides per 100,000 population number from 1991 can't be right, unless
long-time residents remember 1991 as the Year of the Axe-Murderer or something. The overall
homicide rate was 5.8 per 100,000 in 1991, still (or more accurately, again) 5.8 per 100,000 in
2014, and 6.7 per 100,000 in 2015. None of these numbers cover Delaware in glory, but the
overall murder rate bounced up and down between 1991 and 2014 but wound up the same. Then
it spiked up by 15.5% between 2014 and 2015 (or between 1991 and 2015, if you prefer to look
at it that way, but again, small sample size and lots of fluctuation from year to year).

(3) Hard to overemphasize just how subject to small sample bias the homicide rate in Delaware is.
Look at the FBI's UCR table, which goes back to 1960, and you'll see the rate has been as low as
2.3 per 100,000 in 2003 and as high as 10.3 per 100,000 in 1974. It bounces around a lot from
year to year.

(4) While relatively little of the statewide change in murder rate is the City of Wilmington proper,
the murder rate there is still in the 30s per 100,000 residents. That's South Africa and Guatemala
territory.

(5) And most importantly, gun banners can't do math, are sloppy and don't mind lying to push an
agenda. I wouldn't use this study for toilet paper.