Sorry about my earlier reply, I was just curious to see if it was possible to post anything that you wouldn't disagree with or call me ignorant over, I'm glad to see that you do have limits. Though I am struggling to see how it could be interpreted as an insult as someone did
So to respond more fully to your criticism of my informed post.
Quote:
Originally Posted by cebalrai
The above post is utter ignorance. You're just making things up. Total fabrication.
|
No, actually I have a number of qualification in statistics, one being at degree level, as you will see, others since have confirmed.
Quote:
Originally Posted by cebalrai
1) A sample size of 30 or more? Statisticians? Huh? The minimum sample size necessary for validity is a function of how large the group is that's being polled. 30 may be a good enough sample size to measure 80 people. But is it large enough to accurately measure a group of 3,000? 300,000? 3 million? Absolutely not!
|
Population size certainly is relevant, but its a much smaller factor than most people realise, 30 is indeed a pretty good sample for many cases, though I agree it may not be depending on circumstances. However, I have just done the numbers, and for a population size of one million with these poll percentages you only actually need a sample size of 21 to reach 99% confidence in the outcome. 95% is generally accepted as being "statistically significant" and even a smaller sample size is required for that level of confidence. The big factor here is the confidence interval being so massive at 28%.
Quote:
Originally Posted by cebalrai
2) Back to the question of who is voting... It's people that read, and quite possibly post on forums. This is a tiny subset of GW players, not a representative sample across the board. Obviously .001% of GW players are coming here and voting, otherwise we'd have hundreds of thousands of votes on each side.
How many casual (read: don't play/care much) players read forums? How many oblivious new players are here?
How many uber-hour highly invested players are here? Who's more likely to show up here on this bitter, hostile thread (and vote): a new player, a very infrequent player, a female player, young players, or the 25+ hour per week gold farmer?
Unless you're getting a proportional cross-section of all these groups (and more), then this poll is completely invalid horse puckey.
|
So you agree with me that the population here is not representative, as I clearly state
"it can't really represent the true demographic of GW players" who would have though that we could agree on something? There are other factors that should also be considered too, for example, I know of quite a number of players that no longer play GW/visit the forums because the left the game due to the loot scaling.
Quote:
Originally Posted by cebalrai
3) It would not take a "monstrous skew" to overturn 70 to 13 voting. All it takes is polling heavily within one tiny subset that fits a certain profile and you get this invalid bunk.
|
No, you fail to realise how massive a confidence interval of 28% is ((70-13) / 2), this really will take a monstrous skew. While we can identify a number of factors that would make this forum’s demographic to the GW player-base demographic, some skewing one way and some the other, I do not see any difference that would amount to a monstrous skew. I could well be incorrect, that is just my opinion, I respect that you may have a different opinion.
Quote:
Originally Posted by cebalrai
4) Your logic is utter failure when you admit that the poll doesn't represent GW players and then say the results are still meaningful. You're basically saying "okay, so the poll doesn't mean anything, but it still tells us everything we want to know."
You say "this poll is not representative, but it is indicative". All it's indicative of is that 69% of a tiny, non-representative segment of people have a certain opinion.
What the fark? Your logic is atrocious.
|
No, this is precisely why we have things like confidence limits and statistical significance calculation, to put some degree of certainty on things like this. I have calculated the statistical boundaries which identify the scale of skew that can be tolerated. I have identified problems of how representative the sample demographic is and used this logical approach to evaluate the poll. You may not like this approach, but I think it pretty reasonable, there are certainly flaws I would accept, but we have to work with what we have, so I would suggest calling my logic atrocious is a somewhat hysterical outburst.
Quote:
Originally Posted by cebalrai
PS - You're misusing the term "skew". You can't even spell it. "Indicative" isn't used appropriately either. And for the love of Lyssa, "probablistic" isn't a word...
|
1) The use of the word skew is perfectly correct so I have no idea how you consider it misused.
2) Yes, got me on the spelling, it was late, “scew” looks real funny now
3) Indicative is completely appropriate, and indeed is often used as a disclaimer on poll results eg "Results are indicative and may not reflect public opinion - BBC News standard poll disclaimer"
4) Probabilistic (probabilistically) is a perfectly fine, correct and appropriate word, so for the love of Melandru, please try not to get so upset if I missed an "i" in my typing.
A simple question for you. Do you actually believe that the poll results are wrong and that the majority of actual GW players are in favour of the loot scaling? Or was your reply to point out what you perceived as errors in my post? In either event, it could save a lot of time if you could clarify your position on this point.