tzimisce said:You cannot take the percentage based on the the survey sample size to the size of the entire set. Generally speaking, the bigger your sample size, the higher your accuracy becomes, but past a certain threshold, the margin of accuracy becomes good enough for most statistical analyses.
Try this -- Sample Size Calculator - Confidence Level, Confidence Interval, Sample Size, Population Size, Relevant Population - Creative Research Systems
You'll see that to derive a confidence level of 95% with an interval of 5%, you only need 384 responses. To increase confidence level to 99% with a 5% interval, it barely doubles to 666 responses.
(By the way, I'd guesstimate that the volume of iPad sales would be somewhere closer to the 18-20 million mark.)
In our case, we have a strongly biased population ( strongly because we are polling for age and the selection of the participants depends on their age - as I have mentioned, writing on a forum is an activity that is strongly social and cultural and depends intrinsically on the degree of study and on the desire and skills of communication of the polled people. Those are strongle age related). Consequently the numbers polled are (and will probably be) too low in order to offer a good confidence level relative to the complete iPad owners population indifferently how much higher than my 4M guess would be. At least this is my opinion.
Happy (almost) Birthday. You may have put me on your friends list, but I'm still not buying you an iPad2 for your birthday!57 in a few days.
<======== 54 in May, and I have plenty of grey hair to prove it!Shout out for us "old" peeps.
Things are coming full circle. I haven't been carded in over 30 years but now I can't wait to get carded again.... for my senior discount!After all these years I'm finally in the top list,age has its perks and when I find out what they are I will post them.
Al.
1991-C4 said:Well..... lemme see how much I get for my iPad1 on Craigslist!