Username: Password:


Author Topic: Correlation, Causation, Judgement (Hill's Criteria)  (Read 3619 times)

Offline DethPimp

  • Posts: 6846
  • Liked: 1391
  • Country: au
  • Awards Best Writer 2014 Best moderator Best writer Awarded to the user voted Best Member
    • View Profile
    • Awards
Correlation, Causation, Judgement (Hill's Criteria)
« on: February 03, 2013, 05:38:58 PM »
It is true , of course, that correlation does not imply causation.
Classic example: it was thought that polio might be caused by ice cream consumption, as there was a clearly observable and proportional correlation between the two.
This was disproved when it was shown directly that the cause of polio is a virus that is not in any way related to ice cream. The correlation was meaningless, a mere coincidence.
Does this mean that all correlations are meaningless?
If so, it would be bad news. There are many cases, particularly in studies of large populations (eg, epidemiology) where direct, causal links cannot be established and correlations are the only data available.
The classic example here is the link between smoking and cancer. For a start, the cause of cancer as such is not known. In addition,  some smokers never get lung cancer and some non-smokers do.
Does this mean that the question can never be resolved? The medical/legal/scientific consensus is that it does not.
The first part of the solution is to shift from simple cause to risks. Risks can be calculated statistically without stipulating or predicting that an event (carcinogenesis) will be caused in any particular instance.
The second part is to find some way to judge the relevance of statistical correlations in the absence of a verifiable, direct, mechanical causal link.
The medical/scientific/legal communities have accepted that HILL’S CRITERIA provide a sufficiently reliable way of measuring the relevance of correlations in order to guide action in the absence of a demonstrable causal link.

You don’t hear that nearly as often as ‘correlation does not imply causation’, but it is the case. This is unfortunate, because it means that the ‘correlation does not’ slogan becomes an easy way of denying relevant evidence (exactly what the tobacco industry did). A valid principle becomes a form of sophistry.

Typically, what you see is people who act as if they already know the answer selectively using the credo to suggest that nothing can be known  as soon as they hear anything they don't want to know. It's rhetoric posing as epistemology.

Hill’s goal was to show that even causes cannot be ‘known’ in the direct and demonstrable way, there can still be a reasonable basis for action. That is, he does not treat correlation as a debating point, because his goal is not to dominate or silence alternative hypotheses, but as a practical tool. (and let’s remember there is no absolute certainty in any area of science).
The list of the criteria is as follows:
1.   Strength of association (relative risk, odds ratio)
2.   Consistency
3.   Specificity
4.   Temporal relationship (temporality) - not heuristic; factually necessary for cause to precede consequence
5.   Biological gradient (dose-response relationship)
6.   Plausibility (biological plausibility)
7.   Coherence
8.   Experiment (reversibility)
9.   Analogy (consideration of alternate explanations)

For my next post, I will explore the meaning and application of these criteria with reference to an interesting and complex question: the observable correlations between gun control policies and murder rates. Something to look forward to :)

Offline Cheese

  • Administrator
  • *
  • Posts: 7892
  • Liked: 1551
  • Country: us
  • Сука Блять
  • Awards Funniest Member 2014 Meme King 2014
    • View Profile
    • ?????????????
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #1 on: February 03, 2013, 05:54:45 PM »


Too good of a post my good man.

Offline DethPimp

  • Posts: 6846
  • Liked: 1391
  • Country: au
  • Awards Best Writer 2014 Best moderator Best writer Awarded to the user voted Best Member
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #2 on: February 03, 2013, 06:13:30 PM »
<a href="http://www.youtube.com/watch?v=_q_wgw8CsQA" target="_blank">http://www.youtube.com/watch?v=_q_wgw8CsQA</a>

Offline godzilliac

  • Voodoo Baron
  • Posts: 3132
  • Liked: 100
  • Country: ht
    • View Profile
    • Renderosity.com
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #3 on: February 03, 2013, 10:23:35 PM »
The trouble with statistics is that there will always be an element of subjectivity. You could draw perfect valid conclusions based on data that has been collected and analyzed but who's to say if the researchers did not overlook variables that could have been relevant for their survey. And is there a common standard as how to decide which variables are relevant? Think about the very controversial Bell Curve. The researcher(s) intention was to prove a relevant difference in intelligence between white and black people. Several tests were conducted to back up the validity of this statement, yet the general public refuses to acknowledge the outcome of this study.

It all begins with the bias of the researchers and his or her creativity as to what variables they want to include in their study. Statistics are in essence a prediction of chance and not exact science. A high occurrence of correlations might be mistaken for causality which makes matters only more diffuse. I'm not implying that statistics will never be useful but they could very well be manipulated in such a way as to serve a biased proposition in such as a sophisticated way that the conclusion seems valid.

Offline DethPimp

  • Posts: 6846
  • Liked: 1391
  • Country: au
  • Awards Best Writer 2014 Best moderator Best writer Awarded to the user voted Best Member
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #4 on: February 04, 2013, 11:52:01 PM »
Godz, that is exactly what Hill's criteria are designed to guard against. But the statistics I will be using are not subjective. They are verifiable and objective. They don't come from answers to questionnaires, but from hard data.

Offline DethPimp

  • Posts: 6846
  • Liked: 1391
  • Country: au
  • Awards Best Writer 2014 Best moderator Best writer Awarded to the user voted Best Member
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #5 on: February 04, 2013, 11:57:51 PM »
Correlation/Causation/Judgment Pt.2

The correlation we want to explore is that between US gun policy and intentional homicide rates in general (regardless of means).
Comparisons will be restricted to economically developed nations, for reasons that are obvious. There are 24 developed nations as classified by the UN HDI.
To be very clear: even though the United States has by far the highest rate of gun ownership in the entire world, that is not what is being explored here. (although, as we shall, looking at this data only reinforces the evidence that it is gun policy, not numerical prevalence of guns, that correlates with murder rates. That’s the real relevance of Switzerland)
The fact is that the US is the only developed nation in which it is the responsibility of regulating authorities to prove that any individual should be deprived of their right to bear arms, rather than it being the responsibility of the individual to prove they have a valid reason to possess arms. In other words, in the US the onus is on the authorities, not the gun owners. The US may be classified as having a  ‘permissive’ gun policy in these terms (go to  http://www.gunpolicy.org)
So, what we are analyzing here is ease of access, not sheer numbers.
Also, let there be no mistake, I am examining the correlations between gun policy and murder rates in general, NOT just murders using guns. (predictably the correlations there are much, much stronger).




1.   Strength of association (relative risk, odds ratio)
The US rate of intentional  homicide is 4.8.
This is at least more than twice that of any other developed nation, other than Lichtenstein and Luxembourg, at 2.8 and Luembourg at 2.5
However, Lichtenstein can be excluded on the grounds that the population is too low to support meaningful comparisons since it records a rate of 2.8 with just a single murder. This means that any murder at all immediately puts the Lichtenstein rate higher than the US.
Luxembourg is borderline at best in terms of numbers. A mere 12 murders gets a rate of 2.5. But even if we allow this, it is only one point off the halfway mark.
17 developed nations have intentional homicide rates at most a QUARTER of the US (1.2 or under). That is 70 per cent.
So, there is no doubt, the risk of being murdered is significantly higher in the US than any other developed nation. By far. The correlation here is massive. Way higher than is the case for smoking and cancer:
 (Strength of Association. The lung cancer rate for smokers was quite a bit higher than for nonsmokers (e.g., one study estimated that smokers are about 35% more likely than nonsmokers to get lung cancer).
2.   Consistency
There is not a single exception or near exception. The murder rate is far greater in the US than ANY o the 24 comparable nations and the US has by far the most permissive gun laws. The correlation here is ABSOLUTE!
3.   Specificity
How specific is the relationship between intentional homicide and US gun policy?
Switzerland has the third highest rate of gun ownership in the developed world, but a restrictive  gun policy, and a murder close to one eight of the US. If we plot intentional homicide against prevalence of guns the correlation is not nearly so strong.
Even more tellingly, the US does not consistently have higher crimes rates than other developed nations, ONLY in intentional homicides.
So, there is strong evidence that there is a specific link between gun policy and intentional homicides. (if anyone can think of contrasting evidence that the murder rate is more strongly correlate with any other plausible factor, I’d be very interested to see it).
4.   Temporal relationship (temporality) - not heuristic; factually necessary for cause to precede consequence
The ‘correlation does not imply causation’ is above all a caveat against taking this factor as sufficient proof of causation, although it remains a relevant factor.
60 per cent of homicides in the US are committed using guns (this is multiple times higher than any comparable nation). So, the getting of a gun comes before the killing in more than half the murders committed in the US. There is very little evidence that significant numbers of guns are illegally imported. Most murders in the US are committed using guns that were made easily available in the legal market, even if they were illegally diverted at a later point (either by stealing or strawman purchases).
5.   Biological gradient (dose-response relationship)
This is difficult to translate into appropriate terms. It would be necessary to quantify restrictiveness of laws in each of the 24 nations and plot them against murder rates. But: the nation with the most permissive laws does have the highest general murder rate BY FAR.
6.   Plausibility
The argument that the efficiency of guns as a killing tool means that increased availability will cause increased killing is plausible. I defy anyone to suggest it is not.
7.   Coherence
This condition demands that the alleged cause/effect relationship should not contradict known facts about factors relevant to that relationship. The supposition that permissive gun laws might cause a greater murder rate is coherent in this sense (if someone knows any established principle that suggests  there cannot be such a causal relationship. I’ll be glad to hear it).
8.   Experiment (reversibility)
One relevant experiment was the imposition of new gun controls in Australia in 1999. The Australian Bureau of Criminology says that this change correlated with a high drop in murder gun rates,  a significant drop in overall homicide rates, and a COMPLETE STOP to the kind  mass killings which prompted the move. (I’ve already provided sources for this). According to the rhetoric of the US gun lobby, the changes in Australia should have led to more gun murders, more murders in general, and more mass shootings. (they deal with this inconvenient fact by simply lying about the Australian record).
If anyone knows of relevant ‘experiments’ that suggest the opposite, I’ be glad to hear it. You know,  real ones, not ones that someone made up.
9.   Analogy (consideration of alternate explanations)
This is what I’ve been asking people to provide for weeks in post after post. What other explanations are there for the observable and uniquely high US murder rate in relation to ALL other developed nations?
Some things that can be ruled out because they correlate less strongly
Sheer prevalence of guns
 (see Switzerland). Let’s focus on this for a second, because Switzerland (like Australia and Israel) is routinely lied about by the gun lobby. Switzerland has a lot of guns, but does not have permissive gun laws. Follow this link: http://rant-terrific.blogspot.com.au/2013/01/gun-control-part-2-restrictive-vs.html
Overall crime rate (the US does not have higher overall crime rates, just higher murder rates).
Video Games: the same games are played with similar restrictions regarding age in most of the nations in question. There is no correlation here at all.
CONCLUSION: I submit that if no one can offer an alternative explanation that correlates mores strongly with the US murder rate, then US GUN POLICY IS THE MOST LIKELY CAUSE OF THE MASSIVELY HIGHER US MURDER RATE RELATIVE TO COMPARABLE NATIONS.

PS.
   I am not saying, and have repeatedly avoided saying, that this settles the issue. The risk attributable to permissive gun laws is one thing; the VALUE OF THE RIGHT they enshrine is another.
I have never once denied that there is such a value and such a right. It is not something I personally understand, but that doesn’t prove anything. I’ve never owned a gun and I live in a country that does not have a constitutional right to bear arms.
The right to live without constant surveillance, which I do understand and care about, also implies that society is willing to accept a certain level of violence and murder. So, I do accept that some rights are worth such a cost.
So, I am still waiting to hear why people identify their personal freedom with the right to buy guns without checks at shows because of an accidental loophole, for example.
Why does that matter? I don’t get it at all. Tell me.

PPS. It is possible I have not satisfied some of the criteria logically, ie, that I have misunderstood how they apply in this case. I'm open to arguments about that, too.



SOURCES:
http://en.wikipedia.org/wiki/Developed_country
http://en.wikipedia.org/wiki/List_of_countries_by_intentional_homicide_rate
http://www.gunpolicy.org
http://www.southalabama.edu/coe/bset/johnson/bonus/Ch11/Causality%20criteria.pdf
http://rant-terrific.blogspot.com.au/2013/01/gun-control-part-2-restrictive-vs.html


Offline Valjean

  • TWO
  • Administrator
  • *
  • *
  • Posts: 7017
  • Liked: 2457
  • Country: us
  • Writing names on my hollowtips
  • Awards Best admin Best Admin 2014 Most Like John Marston
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #6 on: February 05, 2013, 08:53:25 AM »
RE: your PPS the key fallacy of gun-control laws is that such laws do not actually control guns. They simply disarm law-abiding citizens, while criminals bent on violence still have firearms readily available.  So your actual logical fallacy is you assume you have 1 population when in fact you have 2.  But, I'll even accept this fairly significant... assumption... and include it in my response.

But really Bill you are misusing the word "correlation" to mean " = what OldBill believes."  If they correlated then South Africa... which has ~ 60 million citizens only less than 3 million of them being gun owners.... wouldn't have a murder rate of ... I'm sorry did you say 4.8 for the US?  Let's try on 31.8 for South Africa.

Well I know... those aren't... the developed world.  Those are you... you know... "undevelopeds."  Non-white people so they don't count.  Now Russia on the other hand with a murder rate over the last 10 years between 10 & 20 now that would be confounding given the fact the people were disarmed by the Communists and guns since have been almost totally outlawed including all handguns unless they shoot rubber bullets or are gas pistols.  See that would... it would damage an accusation of correlation.

But I know... that's not quite "the civilized world" ... Russia.  We always considered them the 2nd world.  So let's use.... AMERICA.  Hmmm.  If they correlate then Washington DC with America's lowest gun ownership rate (3.8%) wouldn't also have a murder rate of 17 (positively 3rd world isn't it?)

Saying something correlates doesn't mean that it does.  But, fine.  If its correlation you want, its correlation I shall give you. 
« Last Edit: February 05, 2013, 10:48:23 AM by Valjean »

Offline Valjean

  • TWO
  • Administrator
  • *
  • *
  • Posts: 7017
  • Liked: 2457
  • Country: us
  • Writing names on my hollowtips
  • Awards Best admin Best Admin 2014 Most Like John Marston
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #7 on: February 05, 2013, 12:21:30 PM »
MORE GUNS = LESS MURDERS.  Now there's a correlation we can all put in our pipe and smoke.  (seedy)

Alright Bill, you're claiming huge statistical connections that don't exist.  Now I never liked statistics.  In fact, got a C in Probability & Statistics in my undergrad after failing the final exam, although I did get an A the next 3 times I took statistics so maybe I redeemed myself.  I didn't want to spend my entire life listening to things like " The correlation here is massive. Way higher than is the case for smoking and cancer" when there was no correlation at all.

Correlation
The degree to which two or more attributes or measurements on the same group of elements show a tendency to vary together.


Now how does correlation come into play with gun control.  With gun control there is a negative correlation, meaning where there are the most guns murder rates are the lowest, and where there are the least guns murder rates are the highest.  I will look at guns & crime inside America and as America compares to other nations.  One has to ask the hard questions:

  • Why are gun ownership rates higher in rural areas, but murder higher in urban areas?
  • Why are gun ownership rates highest among whites, but murder rates highest among blacks?
  • How can there exist countries with very high gun ownership rates but very low murder rates if they correlate (Switzerland, Israel, Finland, New Zealand)?
  • How can there exist countries with gun bans or very low gun ownership rates but very high murder rates if they correlate (Mexico, Russia, the ahem.. British Virgin Islands)?
  • Can Europe's low murder rate be attributed to its strict gun control laws?


So here's what I found as it relates to America by itself:

1) The higher gun ownership, the lower the murder rate.
2) The higher gun ownership, the lower your chance of being murdered by a gun.

This is a graph of all 50 states + DC.  Gun ownership % on the bottom.  Y axis is either murders or murders committed by a gun.  The black line shows you that, well, the more guns the lower the murder rate and the fewer murderers used a gun.









As it relates to Europe, since my intellectual compatriot in this discussion, Immanuel Kant, would like to rest on the murder rate in Europe as it relates to America. 

1) The murder rate in Europe was already at an all time low BEFORE gun control laws were enacted.  These gun control measures had no impact on reducing murder. 
2) Once the UK enacted its gun ban, murders went up.  Going to use someone else’s graph for this, one but verified their numbers  here




3) Let’s do one better, then I’m done.  Since Immanuel Kant likes to cite the International Homicide Rate from Wikipedia here it is laid along side Wikipedia’s gun ownership page.  More guns in your country, less murder.  Worldwide.  Period.




Quote from:  Harvard Journal of Law & Public Policy
We assert a political causation for the observed correlation that nations with stringent gun controls tend to have much higher murder rates than nations that allow guns. The political causation is that nations which have violence problems tend to adopt severe gun controls, but these do not reduce violence.


Sources:
Murder Rate in England: Parliament http://www.parliament.uk/documents/commons/lib/research/rp99/rp99-111.pdf
Gallup Poll: http://www.gallup.com/poll/150353/self-reported-gun-ownership-highest-1993.aspx & http://www.statisticbrain.com/gun-ownership-statistics-demographics/
US Bureau of Justice Statistics: http://bjs.ojp.usdoj.gov/index.cfm?ty=tp&tid=43
Crime Rates by State: Center for Disease Control: http://www.cdc.gov/mmwr/pdf/wk/mm6018.pdf  & http://www.infoplease.com/us/statistics/crime-rate-state.html
Gun Ownership rates by state: http://usliberals.about.com/od/Election2012Factors/a/Gun-Owners-As-Percentage-Of-Each-States-Population.htm 
Number of guns per capita: http://en.wikipedia.org/wiki/Number_of_guns_per_capita_by_country
Homicide rate per country:  http://en.wikipedia.org/wiki/List_of_countries_by_intentional_homicide_rate
« Last Edit: February 05, 2013, 12:57:27 PM by Valjean »

Offline booshthelurker

  • Muckraker
  • D WILD BUNCH
  • *
  • Posts: 2965
  • Liked: 483
  • Country: us
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #8 on: February 06, 2013, 02:41:56 AM »
This is one juicy topic!

I offer a 1v1 OldBill vs Valjean debate up for consideration. Both are intelligent and well written, and it would be neat if they had a debate that was moderated. We only need an equal to moderate... I'm at a loss here, and truth be told I genuinely do not think this forum has an impartial third party of their equal. I'm not being sarcastic at all here - it we be awesome IMO.

Just saying... that would be cool. :)

Dirty Wild Sig by DocRikowski
I rep the D the way I do
I can't be you I won't be you
I got soul in me that's older than you
I got my mans with me we colder than you

Offline Valjean

  • TWO
  • Administrator
  • *
  • *
  • Posts: 7017
  • Liked: 2457
  • Country: us
  • Writing names on my hollowtips
  • Awards Best admin Best Admin 2014 Most Like John Marston
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #9 on: February 06, 2013, 07:15:49 AM »
I'm kinda bummed personally :(  I was hoping after a day someone would come back and point out a certain weakness in the data I presented so I can make my final point.  (hint hint)

Offline godzilliac

  • Voodoo Baron
  • Posts: 3132
  • Liked: 100
  • Country: ht
    • View Profile
    • Renderosity.com
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #10 on: February 06, 2013, 07:36:42 AM »
Godz, that is exactly what Hill's criteria are designed to guard against. But the statistics I will be using are not subjective. They are verifiable and objective. They don't come from answers to questionnaires, but from hard data.

Perhaps I didn't make myself clear enough. The statistics may come from hard data but that alone cannot eliminate the preference of the researchers as to which data or variables they choose as to prove their point. For instance, one could compare the number of people in the USA who got prosecuted for the use of marijuana with the number of prosecutions in The Netherlands and draw a valid conclusion based on hard facts that The Netherlands have much less of a drug problem. If you add the fact that use of marijuana is legal in The Netherlands it would drastically change the outcome of the study. If you choose to overlook this fact your data would still be correct but the conclusion might be astray from reality.

Offline Valjean

  • TWO
  • Administrator
  • *
  • *
  • Posts: 7017
  • Liked: 2457
  • Country: us
  • Writing names on my hollowtips
  • Awards Best admin Best Admin 2014 Most Like John Marston
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #11 on: February 06, 2013, 11:00:33 AM »
Perhaps I didn't make myself clear enough. The statistics may come from hard data but that alone cannot eliminate the preference of the researchers as to which data or variables they choose as to prove their point. For instance, one could compare the number of people in the USA who got prosecuted for the use of marijuana with the number of prosecutions in The Netherlands and draw a valid conclusion based on hard facts that The Netherlands have much less of a drug problem. If you add the fact that use of marijuana is legal in The Netherlands it would drastically change the outcome of the study. If you choose to overlook this fact your data would still be correct but the conclusion might be astray from reality.

As long as researchers choose equivalent measures there isn't the type of taint you're describing in results. 

More apples to apples in your example would be length of criminal sentence for possessing an identical drug in America as, say, Netherlands.  The recent Fair Sentencing Act comes to mind for me, especially when you think about how many people we've thrown behind bars for the rest of their life for no crime other than being addicts.  A great many other countries are far less penal.... of course you get caught with 7 ounces of weed in Malaysia you get hung by the neck till dead so I guess some places are worse... But aside from the Muslim countries not many.  /rambling rant

Offline Cheese

  • Administrator
  • *
  • Posts: 7892
  • Liked: 1551
  • Country: us
  • Сука Блять
  • Awards Funniest Member 2014 Meme King 2014
    • View Profile
    • ?????????????
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #12 on: February 06, 2013, 12:34:57 PM »
Are you suggesting high legal gun ownership equals lower murder rates? In that case it would be causation not correlation wouldn't it? If you would say "Illegal guns in the UK is the cause of the increase in murder rates" than that is causation.

Offline Valjean

  • TWO
  • Administrator
  • *
  • *
  • Posts: 7017
  • Liked: 2457
  • Country: us
  • Writing names on my hollowtips
  • Awards Best admin Best Admin 2014 Most Like John Marston
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #13 on: February 06, 2013, 12:47:34 PM »
Are you suggesting high legal gun ownership equals lower murder rates? In that case it would be causation not correlation wouldn't it? If you would say "Illegal guns in the UK is the cause of the increase in murder rates" than that is causation.

Actually causality is a much different standard.  All things causal must correlate, not all things that correlate are causal.  For example... went to my kids kindergarten orientation last night (true story, it was like slow death).  One thing they say is eat a good breakfast, helps them be better learners.  Studies have shown this.  Really though studies have shown they correlate.  Kids who eat a good breakfast do better in school.  That is very different than being able to prove that eating a good breakfast causes you to do well in school.  Cause & effect must harder to prove than remarking that as one variable increases the other also does something (trends upward, downward, etc).

Offline Cheese

  • Administrator
  • *
  • Posts: 7892
  • Liked: 1551
  • Country: us
  • Сука Блять
  • Awards Funniest Member 2014 Meme King 2014
    • View Profile
    • ?????????????
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #14 on: February 06, 2013, 01:47:57 PM »
It has been too long since I analysed well thought out statistics and posts, I need to whip out the metaphorical WD40.

But since you were bummed about your post, I think it sucked because it uses a lot of bold. I hate bold.

Offline Valjean

  • TWO
  • Administrator
  • *
  • *
  • Posts: 7017
  • Liked: 2457
  • Country: us
  • Writing names on my hollowtips
  • Awards Best admin Best Admin 2014 Most Like John Marston
    • View Profile
    • Awards
Re: Correlation, Causation, Judgement (Hill's Criteria)
« Reply #15 on: February 06, 2013, 02:42:51 PM »
I'm neutral on bold.  Personally i hate poodles.  stupid evil poodles and their red eyes.  Oh yeah, and midgets, they freak me out ..