Originally Posted by Lotus
It all depends on the number of people questioned. NPD claimed to have asked tens of thousands of recent HDTV purchasers. Which means their number may or may not be more accurate than yours depending upon the pool of people questioned.
I figure 25% is about right and is large enough of a market for HDM to do well. It's already more popular than LD.
Did you ever hear the joke about the statistician who drowned in a lake with an average depth of only 1"? He stepped into the only 10' deep hole in the lake bed.
The number of people sampled becomes irrelevant once you have a sample size that is STATISTICALLY SIGNIFICANT at either a 95% confidence level or a 99% confidence level.
Once you have a statistically significant sample, the wording of the question and the understanding of precisely what is being asked becomes VERY, VERY important.
The other really important research tool that is also often overlooked is recruitment of research participants from an audience who are good prospects for your product or service. I note this because a good friend of mine who is a vegetarian was recently recruited by NPD for a study they were doing for a large chain of steak houses (Can you say Ruth?): He was NOT EXCLUDED from their survey even after he told them he was a vegetarian. But he did tell them he would eat TOFU steaks, so I guess that "qualified" him for the study.
It's always about the numbers with the big research firms. It is becoming more difficult and more costly to find willing research participants because they either hang up on you when you call them, or they avoid you in the big malls when you try and get them to stop and answer a few questions.
My point here is most of the large research companies now have such a problem getting to that statistically significant sample size they have lowered the requirements for being part of the sample and their client is usually not aware of this.