Originally Posted by twisted_oak
A 6% difference could mean anything, sweet jeebus!
It is not a drastic shift. Give me a break. Open up a new poll and see how it goes a month or so from now. It might swing back the other way.
If this forum represents 10% of the front runners in purchasing tech. A 6% difference of that 10% is nothing. Still ranging at 60+% is a much larger percentage than a 6% variation. That doesn't mean 6% of the total market has shifted.
Hopefully those people who have had the feature mainlined get value from it. If they purchased for the feature alone, I hope there are enough awesome movies to justify that.
I didn't claim the poll was scientific or representative of the general population and would never try to do that extrapolation. I only claimed that you were in denial because you said the numbers were the same.
One thing I learned about poll interpretation in grad school ( scientific data collection and interpretation, UofF Fl SeaGrant) is that it is more important to not derive meaning as an extrapolation of the data when that extrapolation uses other poll data. You may only extrapolate poll data within the domain of the poll itself. ( In layman's lingo, that would mean we can extrapolate the data here within the confines of AVS members) People who extrapolate beyond the domain may be taking a completely scientific collection of data and rendering an unscientific interpretation by contaminating the calculations of the result. Whether or not this is scientific, you still can't execute an unscientific procedure and make an unscientific poll scientific.
In the science of polling, a 6% shift as simple and basic as this was is considered significant
, especially at this early stage of the influx of the technology. In another year we may see the increase be higher but less than 6%. What this data would indicate is not just the direction of the trend, but also the beginnings of data set to determine the RATE of acceptance based on the slope of the curve at the knee.
The 6% shift is NOT what we would call a "slam dunk" as there are only two poll sets to create the curve. So, what we have is a straight line with a slope but not a true rate changing curve over time.
In simple definition of this thread poll and the previous one is:
Of all the respondents who voluntarily chose to vote once in each poll, the trend among AVS respondents was a 6% positive shift in acceptance that 3D is here to stay while resulting in a reduction of 6% from those who feel it is a fad.
It is completely wrong to extrapolate the % in this poll to to any other group of people solely because the AVS represents a unique user group. If you stood on the street corner of Time Square and took the same poll until you obtained the same number of qualified respondents, that too would not be considered scientific and would have no relation to this poll. Hire 100 pollsters to stand on the corner of a busy street in 100 cities and you may get a sense of acceptance of the general population. This would not be a simple math of making the poll bigger as you would restrict the total number of respondents as the same.
Assumptions made in this that could not be made in a general population poll: In order to keep the initial respondents selection similar to the AVS poll the initial assumptions have to be maintained. That would be- Here at AVS the poll assumes a couple of basic truths. 1. It assumes the people here taking the poll are interested in Home theater, audio -video, and film making art and science. 2. It also assumes the respondent knows what 3D is and how it affects the present day industry. So, to make it scientific the public poll would need to have a few qualifier questions that need to be answered in a particular way before the 3 pertinent questions could be qualified for the polling data. The respondent wouldn't know this was being done. e.g. if he said I understand "3D" is what they use on Star Trek Holodeck, then he would be disqualified. If he said present day 3D uses colored glasses, he would be disqualified. These two examples would disqualify because the respondent is basically too ignorant of the technology to render a responsible opinion on the key questions.
Hopefully, this gives you a little clue as to how one way a "Scientific" poll can be constructed.
One thing to watch out for in the Here to stay
numbers is if at some time in the near future this would change positively or negatively. Most good scientific polls eliminate this by doing their sample in a short window of time. This poll does not do that. Like on May 15 all of the sudden the 19.5% jumps to 28% or drops to 5%. Then we would have to look for what caused this level of shift at that time. Why the change now? Did a new invention of glasses free TV make huge news? Did the major TV manufacturers claim there would be no more 2D only sets and all new sets would have at no additional charge 3D feature? Or, did the government announce that 3D was deemed harmful and ordered the industry to end all production? These would be poll changing events, otherwise the shift in the numbers will likely be based on education and commercial availability and will shift positive from the present 19.5% to 22% to 26% to 30% over a longer period of time. In other words, a significant but small amount over time.
So, please understand that a scientists does not need to have the data slap him in the face to see a trend. It just has to be significant and 6% is certainly that in statistical analysis.
Finally, while you can conjecture "what if" that is not based on what the poll says but what your possibilities are. Statistics ( aka polls) are qualified by possible outcomes but measured by probability.