Out For The Morning

Duties call. Send links and general comments. Yes, even you, Guest-known-as-'me.'

Comments

[me]

There is an interesting story at open secrets about how troops deployed abroad have donated 6 to 1 in favor of Obama, the candidate that wants to bring them home from Iraq.

http://www.opensecrets.org/news/2008/08/troops-deployed-abroad-give-61.html

[Casey]

It's late, I'm tired, and I'll track the link later, but...

I'm reminded of the Ronulans who continually claimed that "our troops" were preferentially donating to his campaign by a margin of ...6 to 1...

"me" is obviously uninformed about the grunts. What may loosely and liberally (heh) termed a "conservative" point of view tends show a 3 to 1 split, and that includes "south park" republicans, "leave me the hell alone and I'm an atheist" small-c conservatives, and obnoxious independents.

Even if we reduce that to a 66/33 split, that still leaves a 2 to 1 division with respect to a conservative/"don't tread on me"/independent mindset versus a liberal mindset.

I strongly suspect "me" doesn't follow milbloggers all that closely.

Ok, I lied. Took a second to follow up on the link, and the listings in question cite a whopping ...292... contributions. And that's "from troops deployed overseas at the time of their contributions."

So of all the troops deployed abroad who have actually made campaign contributions, a tremendous 134 soldiers have contributed to Obama, compared to 26 for McCain.

That's a big difference from "troops deployed abroad have donated 6 to 1 in favor of Obama."

According to Wiki, we have:

Iraq 196,600
Germany 57,155
Japan 33,164
South Korea 26,076
Afghanistan 25,700
Italy 9,701
United Kingdom 9,655

Which gives a total of 357,451 serving overseas.

So. Out of 357,451 servicemen/women posted overseas, 134 contributed to Obama, while 26 contributed to John McCain. That's out of a total of 292 recorded contributions.

Let's be nice to "me," and give him all 292 contributions. That gives a result of 8.169e-4, or 0.0008169 of all servicemen/women overseas, or 0.08 of one percent of those serving overseas contributed to Barry.

Another way to put that would be eight one-thousandths of those serving overseas.

Sorry, "me." Nice try, but it's you're embarrassing me, as well as yourself. (heh) Forget "shooting fish in a barrel;" try "dropping daisy-cutters on girl scouts at range zero."

But, hey, keep playing. Maybe you'll win a copy of the home version of our game...

[Tully]

How dare you use actual analysis to mock his propaganda? ;-)

[me]

I presented a link to story that presented an objective fact. The response by Casey, whose laptop apparenty came with a basic calculator, involved grabbing random numbers from the air. e.g.
"What may loosely and liberally (heh) termed a "conservative" point of view tends show a 3 to 1 split...Even if we reduce that to a 66/33 split, that still leaves a 2 to 1 division "

Casey what actual source are you getting these numbers from? Seriously.

Casey in keeping with the thin air theme: despite what you think 3 to 1 and 66/33 are actually very different numbers. So are 3 to 1 and 2 to 1. On top of that none of those 3 correspond to 6:1 which is the proportion of donations Obama has received from military stationed abroad relative to McCain

If the numbers were 6:1 McCain that wouldn't mean anything either right?

Tully: look up the word propaganda.

[notropis]

"If the numbers were 6:1 McCain that wouldn't mean anything either right?"

Right, as long as the total is less than 300 out of more than 300,000, it's completely meaningless, regardless of where they gave their money.

BTW, Obama counts t-shirt, bumper sticker and button purchases as contributions, so you might find a couple hundred troops out of over 300,000 that thought an Obama t-shirt was cool.

[Tully]

I know the definition of the word "propaganda." Apparently you don't, or would prefer that others don't. Isolated information presented without any meaningful and relevant context falls squarely within the definition, and that's exactly what you provided. Casey provided the context you omitted. And Casey cited sources (Wiki AND the link YOU provided, which I up-tracked to verify Casey's numbers. Casey's numers are actually well short of total overseas deployment, as we have smaller troop deployments in LOTS of places, and they add up.)

If you want to play math games, better brush up on the basics. 134 to 26 is 5.15 to 1, not 6 to 1. Don't you ever check your data? Or is it just peachy to assert it without checking as long as it says what you'd like it to? ;-)

Taking it a bit further than Casey did, I note that the figures are from the primary season plus June prelims with no segregation of post-primary donations, and that "Open Secrets" compares apples to oranges in constrasting an open and incomplete partial season to two closed complete ones.

Moving on and doing a little comparison of donations to GOP primary candidates to Dem primary candidates from personnel serving overseas (using YOUR source, mind you) I get 140 contributions to Dem primary candidates versus 152 to GOP primary candidates, a somewhat different ratio.

Since I'm not as kind as Casey and won't allot ALL donations from overseas personnel to Obama but only those that actually went to him, the correct percentage is 0.00037%.

I've had bigger barbecues in my back yard. Come to think of it, I was AT a bigger barbecue of Fort Hood personnel just a little bit ago. Obama's name was not associated with any "positive enthusiasm" the few times I heard it mentioned, though I readily admit they were more concerned with berr, 'cue, the opposite sex, and the next duty roster.

If the numbers were 6:1 McCain that wouldn't mean anything either right?

Not much, at the levels of actual occurence and without a complete election season to compare.

[me]

notropis: "BTW, Obama counts t-shirt, bumper sticker and button purchases as contributions"

me: On those few occasion when it happens so does McCain.

[Tully]

Exactly, notropis. Not enough baseline, and you have to have comprehensive figures for closed seasons from previous campaigns as comparitives. Bottom line from reviewing previous campaign seasons is that very very few deployed troops make donations anyway.

[me]

Tully what was isolated about the information I provided? I gave a link to the story which you took time to read and attempted to find logic flaws in. You then went on to research the data that was sourced in that article. Casey sited and sourced nothing. With all the lies, forgeries, BS intelligence and hero manufacturing during the last 7.5 years I would think that people in your tribe might want to steer clear of using the word propaganda in a public discussion.

What math games? 6:1 has nothing to do with the word/number salad that was presented by your running buddy and neither do yours really. E.g. Comparing donations of the respective parties during the primaries vs the general is irrelevant for multiple reasons. Psychologically donations in the primary are geared more to individual candidates than they are to the candidates party. General election donation are more about the settled party positions influenced by who had been selected to be the nom. If anything the numbers from the general are actually more telling as they are more up to date and accurate. There are many other logic flaws with your premise. The truth is pollsters, exit pollsters and pr people use sampling just like this to predict elections results and it is generally pretty accurate. Rather than misapplying the word propaganda this is the word you really should be thinking about:
http://en.wikipedia.org/wiki/Extrapolation

Ron Paul got more donations from overseas troops than McCain. What possibly could Paul (Republican) and Obama(Democrat) have in common that they would get more donations from troops serving overseas than McCain? Hmmmmm.

[notropis]

One last try, "me," although your total inability to grasp the simplest logic makes it pointless.

Fact 1: Less than 1/10th of 1% of 357,451 servicemen deployed overseas contribute to either major party candidate.

Fact 2: A total of 134 contributed to Obama. You can assume that those 134 support him.

Fact 3: A total of 26 contributed to McCain. You can assume that those 26 support him.

This study implies absolutely nothing about the preferences of the remaining 357,291 people. And there is no basis to extrapolate the proportion 134:26 to the remainder. None. In fact, if we are measuring support purely by financial contributions, we'd say that 99.96% of servicemen overseas have no opinion on who the CIC should be -- which is of course ridiculous.

And, no, McCain doesn't count t-shirt, bumper sticker or button sales as contributions. Neither did Hillary, Romney, Huckabee, or Edwards, for that matter. The idea of counting sales of campaign items as contributions was a new thing with Obama's camp, done to inflate his apparent grassroots support. This was reported in that right-wing rag, the New York Times:

link

"No other campaign is known to have listed paraphernalia sales as donations."

[notropis]

I missed this the first time:

"The truth is pollsters, exit pollsters and pr people use sampling just like this to predict elections results and it is generally pretty accurate."

No, their sampling is absolutely, totally, completely nothing like this. If you're playing over at Wikipedia, you best look up:

"Selection Bias."

This would be an example of "self-selection bias."

(I teach statistics; normally I charge for this; you get it free, this one time only.)

[Casey]

me, yes did to a degree pull that number out of the air, but it comes from general reading and knowledge of discussions with respect to political preferences in the armed forces. The general case seems to be a 70%-80% preference for various forms of "conservative." If you recall, I mentioned several varieties, which explains why I say "conservative" in quotes. I was, in fact, trying to give you the benefit of the doubt by postulating a 66/33 split, although I strongly doubt there are that many soldiers of liberal persuasion serving today.

Complaining that I didn't give cites is like complaining that I guesstimated a 10/10 split between straight and gay men. It's a reasonable first-order approximation from general data and experience.

If you don't like my statement of general differentials, throw out some data showing I'm wrong, instead of whining about it. It shouldn't be that hard to debunk me, no?

And, yes, the article in question had shown McCain up 6-1, I would say the same thing.

The above quibble is secondary to the main issue which Tully underlined it for me. The soldiers who donated while serving overseas are a self-selecting group, and not statistically representative of the armed forces.

BTW, Tully, I think the 6-1 quote came from the original article. I admit I didn't catch that myself.

Whoops. Catching up on following posts. me, me, me... How can you say that "Casey sited and sourced nothing." I actually read, analyzed, and cited your source. I cited and sourced the number of those serving overseas. You are apparently quite ignorant of statistical methods. Since I had to study statistics for a systems analysis degree, I can tell you that it is very necessary to ensure that your sample space is representative of the total population, or your results are nonsense. In this case, the data provided doesn't break down contributions by rank, MOS, or even service. The only cited parameter was "donors contributing more than $200." This means we have no idea how representative the sample space is, nor may we mathematically estimate the minimum samples required for a given confidence interval.

In fact, the original article mentions several caveats against treating the data as representative.

A former West Point professor, Jason Dempsey, noted that the small set of contributions from deployed troops at this point in 2008 -- just 323 donations -- should not be extrapolated to form conclusions about military personnel overall.

Maybe me didn't read that far. :)

Ron Paul got more donations from overseas troops than McCain. What possibly could Paul (Republican) and Obama(Democrat) have in common that they would get more donations from troops serving overseas than McCain?

Um, they're both popular with the wackjobs?

You did ask... ;)

[Tully]

ROFLMAO. Casey, notropis, don't you just love strident innumeracy combined with clumsy attempts at employing logical fallacies? But any valid analysis cuts right through it.

BTW, Tully, I think the 6-1 quote came from the original article. I admit I didn't catch that myself.

Both from Me's first entry and the linked article. One glance at that ratio and I knew it wasn't 6 to 1 as claimed, but barely over 5 to 1. 6 times 26 is 156, not 134. And 5 times 26 is 130. For anyone used to quickly figuring ratios no time elapsed at all, as the discrepancy shouts out. Not exactly rocket science.

Me presented the numbers but didn't even look at them close enough to spot the obvious. (Or didn't want to mention the obvious--your own mileage there.) In any case the ratio was obviously wrong, at a mere glance. Me then tried to avoid any responsibility for the numbers or the propagandic import in the presentation thereof by playing the "only the messenger" card and then trying to redefine propaganda. Ipse dixit! One cannot defend and disclaim at the same time without engaging the particulars--and me cannot win that one on the particulars.

Even if we had a full season (complete set) of data, it would only be good for comparing to previous season's data sets, which as both of you quite ably pointed out would tell us nothing at all about the larger population universe of deployed troops, but only about the particular self-selected sample itself as compared to the particular self-selected samples of previous cycles. Apples to apples.

As for counting "item purchases" as campaign donations, it's not completely original to Obama, though yeah, a sale is not a verified assurance of actual support. I have a talking Kinky Freidman doll action figure I obtained in the last Texas gubernatorial election as a "gift" for my "donation" to his campaign--a donation I made specifically to get the doll action figure. And I don't even live in Texas. I don't KNOW that they reporteed those de facto sales as "donations," but they were promoted as "gifts" for certain levels of donations.

I still laugh every time that doll action figure says "Why the hell not?" and "How hard could it be?"

[me]

notropis: sorry for your students if indeed you do teach statistics but yes extrapolation i.e. sampling is how polling works in the real world. The daily general election poll numbers are not the result of every registered voter being asked who they support that day. A sampling (the fraction of the whole)of voters are polled and that number is extrapolated to represent the larger population.

Your facts are indeed facts but they do nothing for your argument one way or the other. Throwing out the term "selection bias" is a red herring too. Selection bias is when, for example, Fox news conducts "a poll" in which they ask their +90% conservative audience if they think x,y,z and the results are presented as a result of what all Americans think. In the case of what is being discussed, the fact that Obama has received a larger amount of donations from servicemen stationed abroad at a ratio of 6:1 compared to McCain is a fact.

These numbers are taken very seriously by pollsters and pr people because they reflect the level of excitement, commitment and numbers for a given candidate. When companies prepare add buys for their products these are precisely the types of numbers they would look at to determine the success of their product relative to the competing brand. And if you want to bring in data about overseas donations from previous elections I welcome it as I suspect it would support my argument even more.

Casey:
It wasn't just that you didn't site anything it was that you threw out a bunch of random numbers, said they were related to each other and implied these figures were based in facts. Because the figures did not convert to what you said they did I was curious what standards you were using to relate these random numbers. Using your own a-priori experience as basis for your facts is an example of the selection bias that your running buddy prof(?) notropis is talking about. Self selection has nothing to do with what I brought up though. As I spelled out before these numbers indicate the quantity and quality of excitement as well commitment for a candidate. These numbers could also reasonable be used to project turnout.

Casey:"Um, they're both popular with the wackjobs?"

Maybe but I think it is more plausible that people overseas are just not that excited by McCain, GWB or the Republican party.

I could quible over semantics about the keychain/t-shirt meme or how you guys only seem to find logic flaws in the facts presented in "teh MSM" when it doesn't support your view %110 but give links to it when is suits you. Suffice it to say where exactly do you think the money goes when people buy the McCain polo shirts($50), McCain ice scrapers($10) and the $200 McCain nautical lapel pins for sale at the McCain campaign web store? I would love if you could explain that.

Have you seen the heard about the numbers of voters that will no longer admit their Republican party affiliation? Or have you seen the list of major Republicans up for re-election that have declined to attend the Republican convention because they don't want to be seen there on camera? This is probably related to the same underlying reason Obama got more donations than McCain at a ratio of 6:1.

[me]

Tully The pink elephant in your post is 5:1 is still 5 time more than McCain.

[Tully]

Ah, more strident innumeracy and open propagandizing, complete to the requisite repetitive noisy idiotic ignorance! Apparently me feels that if me says it often enough it magically becomes true!

Me, you know absolutely NOTHING (or at least nothing correct) about the mechanics and methodology of polling and valid sampling technique. The figures you cited are NOT A POLLING SAMPLE AT ALL, and being self-selected are decidedly not random. Notropis is dead on, and you're proving in copious detail that you haven't a clue about even the basics. There's a game show waiting for you. Gotten your callback from Foxworthy yet? :-D

These numbers are taken very seriously by pollsters

No, the numbers you cited are only taken seriously by idiots grasping at anything. Pollsters know better. I've been one, and work with professional pollsters fairly continuously in real-world elections. For candidates from both sides of the Great Divide.

I suspect it would support my argument even more.

Yet in trying to dodge the propaganda point you said you weren't making an argument, just pointing out something you thought of itnerest. Thank you for confirming your obvious lie as also being a blatant one, and for openly exhibiting the deeper content of your character, thereby clearing up any lingering doubts.

These numbers could also reasonable be used to project turnout.

Only in Bizarro World, where flying unicorns fart pretty rainbows across the violet sky. But if you believe so, go right ahead and project. Cobble together your projection. DOn't forget to justify all data used AND thoroughly document your methodology for review. That would be amusing. I suspect we could have endless fun with it, given your demonstrated levels of ignorance regarding mathematics, polling, and statistical methodology.

Gee, me's almost as much fun as the Paulbots. Just not quite as creative or intelligent.

[notropis]

Tully,

Thanks. I wasn't going to bother, and now I don't have to. It's pretty much hopeless, isn't it?

[Tully]

Yep. Unless you just enjoy the exercise and can spare the time.

[me]

Tullly you are a total drama queen. And just like any experienced drama queen you juggle between the rhetoric of the expert and flamboyant pyrotechnics whose sole purpose is to divert attention. So stamp your 6 inch strappy heeled feet and snap your painted fingers at me all you want. You and prof(?) notropolis are simply putting on a show.

Let us be clear. I gave a link to an article that stated that Obama was getting more donations than McCain from servicemen stationed abroad at a ratio of 6:1. The immediate response was that this number was wrong and a set of seemingly random numbers were put forward as more accurate figures.

After seeing these random numbers and inquiring about where they came from I got responses that basically alternated between “because I am an expert” and “look at that shiny thing over there!”

Yes I asked for some references. But I don’t think “guys I know”, “I read some stuff”, “because I thought about it” and “I looked it up on Wikipedia” counts. Yes I agree Wikipedia is a relatively reliable source of facts for things like this that is why it is important to be specific about what it is you looked up.

I was given numbers about troop levels in various countries, a statement that Obama counts t-shirt sales as a donation while McCain doesn’t (yet no explanation for where the official McCain campaign ice scrapper money for sale at the McCain store ends up), accusations of being a propagandist, mentions of self selection and statements that these numbers are not truly representative enough. And despite this you people took the time to tell me what the “real” ratios of military support is yet you all based these figures on much more suspect sources than what was mentioned in the article -re:“based on guys I know and stuff I read” followed by the classic "this guys just don't get it".

Yes 6:1 is incorrect and I suspect the editor of the article rounded the number because it sounds cleaner as a title than both the 5.1 to 1 ratio or the 5 to 1.6 ratio that Tully alternately calculated as the one true ratio. Yet it still does not explain where the random numbers came from. I think if you can take time to calculate the rounding error in the 6:1 number than surely you can tell me how specifically the magic numbers were arrived at.

Most of you at some point resorted to claiming professional expertise as if this glorified wig flicking would magically give meaning to those random magical numbers. Coincidentally I have some insight into academia and the world of political consulting. No serious academic would resort to announcing their employment status to win a disagreement. The only people who would do that are liars and adjuncts for small town community colleges. Prof(?)Notroplis which of these are you? If I was guaranteed I could collect on it I would bet my house on you being one of those two.

Tully’s accusation of me avoiding something is text book right wing drama queening No specifics no details just a general “he is avoiding the issue”. It works on the rubes because you don’t have to give specifics that can be openly debated and it carries the insinuation of cowardice. Tully will you avoid answering this specific question? What exactly did I avoid in my previous exchanges? I'll be here if you decide to walk the walk Miss Thing.

Tully it is pretty obvious that in whatever niche you inhabit you are used to people not calling you on your drama queening. I suspect they are too stupid, too tired, too immersed to tell you that your artifice is way overwrought. Either that or your fellow travelers go along with your pantomimes because they are looking for any flimsy straw to make their backwardness righteous. But just because your drama queening impresses your mother and the people in the checkout line at Target does not make it anything more than just a performance.

[Tully]

LMAO. Sounds like I hit squarely on a sore spot.

You and prof(?) notropolis are simply putting on a show.

Yep. The show of demonstrating your (potentially intentional) ignorance of polling and statistical methodology, your inability to construct a coherent argument using facts instead of logical fallacies, and your obvious intent to propagandize.

I answered your question before you even asked it--you were propagandizing and trying to avoid admitting it. And pretty much everything you keep ranting about in your projective ad hominem screed has already been addressed above. You keep waving that wand hoping it will make everything you want to believe or assert true. It's still not working.

Do let us know when you actually manage a projection using your own unique methods.

[notropis]

Well, if you’re banking on “liar,” you’d lose a house. And if the operative word is “adjunct,” you’d still be out of luck. But small-town, yep. Community college, check. Don’t know what either of those has to do with anything – you’re throwing them out as some apparent insult or put-down. But we’re fully accredited, and I’m full-time, tenured, 20+ years teaching experience, along with quite a bit of statistical consulting work with both public and private organizations over the years, none of which has anything to do with what I told you, above, since it doesn’t take any expertise, or even any math, to understand the idea of “self-selection bias.” I just mentioned the fact that I teach statistics to show how lucky you were to be getting a free lesson, when others have to pay a couple hundred bucks a credit hour (we're still cheap!)

I simply pointed out the facts to you. The numbers you cite are not a statistical survey, they’re an analysis of a tiny group of highly self-selected individuals (turns out, to remove the “Obama t-shirt effect” – which I quoted from the New York Times – they limited their study to contributions of $200 or more, making extrapolation LESS valid, since the higher the level of self-selection, the greater the possibility for bias.) Consequently, extrapolating the operative characteristic (whom they support) to the larger population isn’t statistically valid. That’s a simple fact. It’s not arguable. You learn it in freshman level statistics – we teach it week 2 in MATH 1105, also available on-line (not taught by me, this semester, don’t worry; we have a PhD from UNC with a dozen years experience doing research for a biotech firm teaching it; you should find her, at least, credible), should you care to sign up and learn something.

In the meantime, you’ve chosen not to believe it, and learn something, and instead do a lot of name-calling and juvenile taunting; so that ends the conversation, such as it was.

[me]

and I simply asked where the magic numbers came from