Reply #64

In the discussion thread: Take this you gun nutters !!! [View all]

Fri Dec 21, 2012, 11:59 AM

jody (26,624 posts)

64. First, OP regression line is not supported. Second regression models do not prove cause. Third I

asked the OP author "2. How do you explain the DoJ report below when all have same access to the guns you cite?"

If OP presents the statistical results supporting the regression and the R square is "empirical proof" that the "rate of guns to population" predicts the "rate of gun deaths to population", then the DoJ graphs I presented would not occur.

That's easy to understand because both sub populations in the DoJ data had the exact same "rate of guns to population" but the"rate of gun deaths to population" are seven times different.

Either the DoJ data is wrong or the OP's model is not the "empirical proof" the author claimed.

Reply to this post

Back to OP Alert abuse Link to post in-thread

Always highlight: 10 newest replies | Replies posted after I mark a forum

Replies to this discussion thread

88 replies	Author	Time	Post
Take this you gun nutters !!! [View all]	srican69	Dec 2012	OP
Oh my gosh! I'd be safer in Chile!	tblue	Dec 2012	#1
We're not off the rails.	ehrenfeucht games	Dec 2012	#5
What on earth are you suggesting?	EOTE	Dec 2012	#8
If I am reading the chart correctly, we are about where...	ehrenfeucht games	Dec 2012	#10
The point of the graph is that the more guns you have, the more gun deaths there are.	EOTE	Dec 2012	#12
Whether we need more or fewer guns depends on whether we want more or fewer gun deaths.	ehrenfeucht games	Dec 2012	#15
You needn't need be so subtle.	EOTE	Dec 2012	#17
I've been called a lot of things, but "subtle" is a new one.	ehrenfeucht games	Dec 2012	#21
Glad I could be the first :p NT	EOTE	Dec 2012	#61
BTW, the initial post also makes the point that we are NOT an outlier on that chart. (nt)	ehrenfeucht games	Dec 2012	#11
I posted the op...an outlier in statistics is a datapoint	srican69	Dec 2012	#14
I agree. (nt)	ehrenfeucht games	Dec 2012	#16
Over half of them are suicides	Mojorabbit	Dec 2012	#84
you seem to be reading the chart wrong	Motown_Johnny	Dec 2012	#86
you would be safer in any of the countries in the analysis above	srican69	Dec 2012	#7
You made me lol... hehe.	OneTenthofOnePercent	Dec 2012	#48
How do you explain the DoJ report below when all have same access to the guns you cite?	jody	Dec 2012	#2
Are you saying the info in the OP is wrong?	morningfog	Dec 2012	#3
Here is one for children:	morningfog	Dec 2012	#4
you are doing timeseries in USA...I am doing a snapshot analysis	srican69	Dec 2012	#6
No, it's not so subtly injecting race into the discussion	Fumesucker	Dec 2012	#18
Ridiculous. You can't use the race card to avoid DoJ's report. Clearly it's not just guns as the OP	jody	Dec 2012	#23
You were the one that played the race card	Fumesucker	Dec 2012	#32
OP posted a regression model and claimed guns caused homicides. Is there anyone other than me who	jody	Dec 2012	#34
The US seems to be about average on the chart the OP put up	Fumesucker	Dec 2012	#37
Fumesucker perhaps you might think I'm your enemy, other DUers do, but I'm not.	jody	Dec 2012	#40
I don't have a problem with you, we're fine	Fumesucker	Dec 2012	#44
Simple fact, more guns cause higher homicide rates.	Spider Jerusalem	Dec 2012	#38
Please explain how that supports the statistical credibility of the OP's model. nt	jody	Dec 2012	#41
How the fact that more guns means more crime in areas with similar social/socioeconomic factors?	Spider Jerusalem	Dec 2012	#43
Or perhaps you're too stupid to explain. So we've exchanged insults. Can you answer the question?	jody	Dec 2012	#45
Since you evidently ARE too stupid to grasp the essential point:	Spider Jerusalem	Dec 2012	#46
Have a blissful evening and goodbye. nt	jody	Dec 2012	#47
That's DoJ's report, not mine. You asserted guns cause homicide and I asked why the DoJ data	jody	Dec 2012	#22
doesn't matter if it is your analysis or DOJ..your graph shows that	srican69	Dec 2012	#26
LOL why don't you provide all the stats for your linear regression because that's what your OP	jody	Dec 2012	#31
Your little graph says nothing to contradict the OP.	EOTE	Dec 2012	#9
The DoJ data shows homicide rates about seven times different when both sub populations have	jody	Dec 2012	#25
I don't understand a word of this post and I don't think you do either.	yardwork	Dec 2012	#51
I'm not responsible for you being unable to understand simple concepts. nt	jody	Dec 2012	#54
Christ, this is not difficult to understand.	EOTE	Dec 2012	#62
OP says "Here is empirical proof on the effect of guns" OP says it's a regression model but such	jody	Dec 2012	#63
Gun related deaths vs. homicides	rucky	Dec 2012	#87
Ask the OP author. nt	jody	Dec 2012	#88
Why would you need to. It's not a comparable statistic. One is a snapshot. The other is a linear.	Flabbergasted	Dec 2012	#49
Both deal with cause of homicide. Sorry you don't understand. nt	jody	Dec 2012	#56
Your graph does not refute the op. it just doesn't. If it does clearly share your reasoning. Nt	Flabbergasted	Dec 2012	#60
First, OP regression line is not supported. Second regression models do not prove cause. Third I	jody	Dec 2012	#64
Strawman. Nt	Flabbergasted	Dec 2012	#67
Straw-man to anyone who doesn't understand the simplest concepts of regression models. nt	jody	Dec 2012	#68
Strawman:	Flabbergasted	Dec 2012	#69
Your chart doesn't refute the OP's chart. They are two completely separate issues.	yardwork	Dec 2012	#50
No they both deal with cause of homicide. Sorry you don't understand. nt	jody	Dec 2012	#55
Stop being an idiot.	baldguy	Dec 2012	#59
OP submitted a line through a scatter diagram. Said it was a regression and "empirical proof on the	jody	Dec 2012	#65
not to be overly technical but..	srican69	Dec 2012	#13
I think I can speak for the DU community here	Fumesucker	Dec 2012	#19
You don't speak for me. OP makes assertions that are not supported by the model as presented. nt	jody	Dec 2012	#35
Please provide all the statistics for your model so one can determine the efficacy of your claim. nt	jody	Dec 2012	#28
NRA: "Only More Guns Will Solve The Problem Of Gun Violence"	Forrestted	Dec 2012	#20
Where's Brazil?	k2qb3	Dec 2012	#24
what next...why don't you throw in natural deaths as well	srican69	Dec 2012	#30
Thank you for the recommendation...	k2qb3	Dec 2012	#83
I would like to see a chart of how many deaths by "illegal guns", which	crazyjoe	Dec 2012	#27
Who manufactures illegal guns these days?	ehrenfeucht games	Dec 2012	#53
deaths by people whom are in possession of a gun illegally, ok?	crazyjoe	Dec 2012	#85
Our residual is even higher than the Banana Republics	Doctor_J	Dec 2012	#29
What the chart says -	socialindependocrat	Dec 2012	#33
Culture is certainly a factor. Perhaps that's why Japan with low gun rates compared to the US has a	jody	Dec 2012	#36
I think that everyone agrees that people who kill others are mentally unstable.	socialindependocrat	Dec 2012	#39
"mentally unstable" that's not my field but that gets my lay-person's vote for a primary	jody	Dec 2012	#42
And the propaganda king returns for another round	DainBramaged	Dec 2012	#77
The handful of armed people in Barbados are nothing to fuck with	cthulu2016	Dec 2012	#52
Ah, so you include suicides in that number?	krispos42	Dec 2012	#57
Bill Hicks - "You'd be a fool and a communist to make a connection between having a gun..."	Initech	Dec 2012	#58
srican69 what's the R square for your "regression line"? nt	jody	Dec 2012	#66
adjusted R -square of 0.85	srican69	Dec 2012	#70
Thanks. FBI data for the US is at variance with the data you used for your analysis.	jody	Dec 2012	#74
this includes gun related suicides as well as homicides	srican69	Dec 2012	#76
If your data includes suicides, then IMO your model is fatally flawed because it excludes Japan	jody	Dec 2012	#78
it had to be a gun related suicide to count	srican69	Dec 2012	#79
Doesn't matter you cooked the data. You and I know if you presented that at a local chapter of	jody	Dec 2012	#80
congrats on graduating from the Jack Welch school of criticism	srican69	Dec 2012	#81
That's the response I expected given the model you concocted. nt	jody	Dec 2012	#82
BTW The R-square does not matter - you need to look at	srican69	Dec 2012	#71
We disagree. R square in your model can be interpreted as the percent of variation in the dependent	jody	Dec 2012	#75
Source? Link?	guardian	Dec 2012	#72
The analysis is mine .. source data is from wikipedia filtered to OECD countries	srican69	Dec 2012	#73

Edit History

Please login to view edit histories.