"You're one man?"

"Now we are two."

He is not saying that their answers are not correct. He is saying that you cannot base a full opinion on the opinion of the minority. You got answersa out of 31 people. How many people play Heroclix? By your estimates and "results" there must only be 31 people.

Perhaps you're not firmiliar with the mathmatics behind the calculations of survey accuracy. Here I'll just let the web site speak for itself, it will no doubt describe itself better than I will:

This Sample Size Calculator is presented as a public service of Creative Research Systems. You can use it to determine how many people you need to interview in order to get results that reflect the target population as precisely as needed. You can also find the level of precision you have in an existing sample.

I used a target player base of 200,000 players...I don't know if that's correct, but so far as the formula goes once you get above a certain number the actual number has very little bearing on the results. I used a confidence level of 95%, and the resulting confidence interval was plus or minus 8.68%. The sample size used to attain that interval was 31 people.

Or in laymen's terms: 93% plus or minus 8.68+ of the polled demographic does not think powercreep affects their purchases. (That statement has a 95% chance of being true. If you up the interval to 11.5 % then it has a 99% chance of being true.)

(Notice I said "polled demographic" - Naysayers in this thread are still welcome to argue that the HCRealms community is not an accurate demographic...and I suppose you're welcome to argue that the answers given in this poll were affected by bias...but there doesn't seem to be any evedence to support that claim.)

That one was mine. So is "pandas." Both were requested by someone in this thread (you?), and I saw no reason not to oblige...

And as for the opinions of 31 people reflecting a hypothetical 200,000 Heroclix players... what the heck? Brazil, I really don't think that those numbers would yield 93% accuracy or better (but to be honest, I don't remember everything from prob/stat in high school). Seriously, less than one percent of Heroclix players represents the whole? Especially when considering the bias of anyone who volunteers themselves by voting in the poll, that data is anything but reflective of the whole.

But seriously. Am I expected to believe that .0155% of Heroclix players can speak for the other 99.9845%? I'm not saying that you're wrong, but your poll doesn't provide much credibility to your argument.

But there is no way to calculate it correctly unless you have all 200,000 players take the poll. Guess-ti-mating what you think they will say is just not probable.

Why not actually make this an actual poll? Scrolling through tons of pages and arguements to try to find substantial answers just makes me feel like I am banging my head on the wall.

Yeah, that's just the math of it. Here's the link to the site I used, but there are many that can calculate this and you can probably even dig up a proof of the formula used with a bit of digging. (http://www.surveysystem.com/sscalc.htm)

So the way it works as far as sample size goes, once your "Total population you're trying to represent reaches a certain point, it gets to a point where it's pretty much "as bad as it can get"...so the numbers really don't change but a few decimals if I said the total player base was 2000 or 2 million...the accuracy stays more or less the same. I know it "Seems funny" but that's the way the probabilities work out.

I didn't say 100% "correctly calculated". The probabilities I quoted indicated that there is a 95% chance of "correctness" that the actual "Correct" percentage, if all 200,000 players were surveyed would fall within 8.68 percent of 93.5% value found from the 31 people I did survey. Furthermore if you widen out that error chance to plus or minus 11.5% then there is a 99% chance the correct number is in that range.

But you are correct you won't attain 100% certainty, until you poll nearly all of the 200,000.... but I can live with the 95% chance it's within 8.68%.

Brazil is saying that his poll is accurate with a 9% margin

Hence his conclusion is that Powercreep has no effect on the purchases of 93% +/-9% of people who play Heroclix. Hence it could be anywhere from 84% to 100%.

This is of course assuming he used a random sample, which no one seriously thinks he did.


I'm done trying to convince anyone on this thread that people are in general, bad judges of their own motivations. Studies published all over Social Science Journals have repeatedly demonstrated this point. Stating that it doesn't apply to you is simply a conceit.

If nothing else, this thread has clearly proven the bolded part.

For what it's worth, I would be considered a completist by many people. I try to get every character in a set, within reason. I won't spend through the nose for chase figures, but I have just completed M&M (including LEs), I've got all of Sinister (all REVUs) and Supernova (all LEs exept zombies), and I'm no more than 3-4 figures shy of completing every other set since... Collateral Damage or Armour Wars, I think.

Possibly! I often feel the need to be a "voice of reason" when I see people write things that are needlessly negative; it's entirely likely that some part of me viewed your initial questions as having a negative bias and the way that I answered was affected in reaction to that. I can admit to my biases.

It doesn't matter whether any bias supports or undermines the position that you (consciously or unconsciously) want to promote; it's still a bias, and invalidates your "research" as anything more than a vox pop. There's nothing wrong with that - anecdotal experiences can be very interesting to read - but it's really not research, and shouldn't be dressed up as such.

I am certainly tempted to keep them in mind for future threads where they may also be relevant.

Man, you are such a WK stooge. My venue likes to tag, and that's the way we're going to stay! What are you gonna do, report us?

But is there any compelling reason to think that the demographic who participated in this poll would be that much different than other heroclix players? I would actually think that "casual players" who most likely don't hang out on HCRealms, would be even less likely to notice powercreep, much less have it affect their purchasing habits. Hence I would guess polling them would porduce a larger percentage who was unaffected. However, I will say that ... it was either you or Zero_Cochrane who brought up a good point about there being a lower percentage of collectors in the non-hcrealms community and the collectors purchasing habits are set irregardless of powercreep....so that may sway things the other way. That was a reasonable reason for a possible shift the other way. Still the responce was so overwelmingly that powercreep had no effect, even amungst the non-collectors who answered here, that I doubt the numbers would be that much different.

If Social Sciences have demonstrated conclusively that people are bad judges of their own motivations.....then why are their polls at all? Wouldn't EVERY poll be bogus if people who don't know thier own opinion are asked to give it, since you said that is a topic Social Sciences have shown they know nothing about?

Why to public opinion polls continue to exist if that's the case?

And why are you bothering to try to answer this question? You're just going to give your opinion anyway, which you've said yourself you're not capable of knowing?

What arguement am I going to face next? The Chewbacca defense?

Seriously, have you ever seen an angry Wookie? You DO NOT want to make him mad.

Question 1> Do you like Powercreep? Yes

Question 2> How does powercreep affect the amount you spend on new sets? Spend the same

Discussion> If there was no powercreep what do have to look forward too. Another Spiderman that just does more of the same with twist?

But the creep needs to be slow

It's not that they know nothing about it, it's that most people are unaware of all the motivations that go into their decisions. People's self perceptions often Bias their responses against their actual motivations.

1) Most Opinion Polls begin "X percent of people say/claim they believe Y' which is vastly different than "X percent of beople believe Y".
2) Polling opinions is a complex process requiring a lot of study. In the University of California Research Methods in Psych is a required course for upper level Psych, and about 1/3 -1/2 of that course (depending on where you take it) deals with polling. In order to actually run a psychologically valid poll you need several more courses in advanced statistics and polling techniques.

This is kind of one of the US's problems with the social sciences: Loads of people are curious about it, but not curious enough to actually understand
the involved complexities.

Public opinion polls are janky, not because the results are wrong, but because most people, the media especially, don't have the technical expertise to correctly interpret the results.

I'm referring to scientific methods of polling. My opinion doesn't matter. I could believe the sky is blood red, but that doesn't make it true. Science requires opinions in formulating experiments and interpreting result for direction in further experimentation, not in reporting the results.

However you are right in the sense that people who say they would answer the same in my pseudo-poll may or may not be protecting their own self image, and hence be mistaken. They are certainly biased by the presentation of the poll.
Brainwashing technique no 3, Keeping you in line with shame
as well as
Straw man argument

You missed the reference. The "Chewbacca Defense" was made famous in an episode of South Park entitled "Chef Aid" which depicted it as the sole means of defense offered by Johnny Cochran while defending O.J.Simpson. (I highly recommend more research in the form of watching this episode if you haven't already seen it.)