I think we're probably a pretty good sample of the overall kiting population.

It's a free forum - so no income bias
It's an Internet forum - so the media of delivery is generally used by both younger (school age) and older (working age) populations
It's an international forum, so no national/cultural bias
It's not single-vendor sponsored - so it's not biased towards one manufacturer's audience or marketing image
It's a popular forum that is well known - so it's not biased by obscurity
It's got Toby as a web master - so it has a reasonable level of tolerance with pretty minimal content controls
It's got lots of pictures of Susi Mai and that helps.

I 'd be willing to bet that this is the one of the most "typical" samplings of the kiting population that you would find at any gathering.

There are about 25,000 members on this forum, representing a very small fraction of actual kiteboarders.

There are 400 votes cast as of this writing, representing only about 1.6% of forum members.

For every single member that cast a vote, more than 60 members and thousands of other kiteboarders didn't bother, know or care.

Suppose that one out of every 500 kiteboarders under age 25 knows about this forum, is a member of this forum, and cares enough to cast a vote. Now suppose that only one in 5,000 kiteboarders over age 50 does the same. That's very plausible, and it would skew the results by a factor of 10.

Online activity is a cultural thing. Don't imagine for a second that people from different age groups and other cultural sub-groups do it in equal proportions. That's why online polling of any kind is normally of little use other than hype and entertainment.

Phone polls aren't much better. For example, voters who are more symapthetic to liberal government policies may also be more sympathetic to giving the time of day to telephone pollsters that interrupt their dinner.

Cheers,
James

Those are all considerations that are already incorporated into statistics. If your sample is sufficiently random and is close to the distribution of the population, then it takes a surprising small sample size to generate a high confidence level for the conclusion.

As far as I can tell, this forum would appear to match the general distribution of the whole kiting population. In that case, 400 samples provide a pretty high level of confidence for age as a single variable. (I'd have to go look it up in the tables. I thought it was something like a sample size of 400 provide a 5% margin of error at a 95% confidence level.) According to an online sample size calculator, if you want to run a survey with a 95% confidence level assuming a 5% margin of error against a population of 1,000,000.... you only need a sample size of 384.

http://www.surveysystem.com/sscalc.htm

here's another one:

http://www.raosoft.com/samplesize.html

Again, what's critical is how close is the sample matches the distribution (randomness) of the actual population. I would still opine that the members here are about as close to a good representation of the overall kiting population as you can find anywhere.

That may be so, for a random sample. But this isn't a random sample. It's a non-random, self-selected sample.

With respect intended, I don't think you're quite grasping this. It's true that most kiters have some access to this forum, and many from all or most age groups participate to some degree. But the probability of kiters from different age groups participating in very different proportions is possible, and I suspect almost a certainty. Just because you know folks in their 60's that post to this forum, you mustn't assume for a second that as high a proportion of 60+ folks post, compared to 30- folks.

Best regards,
James

Good thoughts back at'cha! Good conversation.

So... you think that the older age groups might be underrepresented?

I would hypothesize that the 50+ age group of kiters is more likely to use the Internet and participate in on-line forums than the average 50+ age group of the world. (Great... another hypothesis.... another study... but we definitely cannot use this forum as a basis for our sample.... It would come out 100% because there are no non-Internet users using this forum...) The kiter population appears to be more outgoing, social and still-active than the average older non-kiter population. I think this is true of any group of people with a heart-felt hobby - whether it be kiting, biking, running, gourd-carving or song writing. Within the context of avid hobbyists, I think there is less difference in age groups as far as participation in Internet forums goes.

I agree that challenging the randomness of the sample is always a fair topic in statistics. Given that this is a "stock argument" that can be levied against any study, I vote to turn our attention to the available groups from which to draw our sample. Who would you choose as a random sample? A mail-out survey from some magazine's mailing list? The crowd at the Brazil PKRA competition? A couple of locale-specific beaches across the world? Besides the strength of statistics on its own, I would claim that the sample from this forum is one of the most "random" of any that are available.

Yes, likely.


...or the least "unrandom" might be a better way to put it. I agree with you, and none of the examples are random. A fairly random study could be done, but I don't think anyone cares enough to go to the trouble and expense, so we can all just carry on sticking with our hunches.

Cheers,
James

to confirm the poll, just do this: go to the beach and look at the riders!

In Cumbuco the majority is exactly what the poll says. Add few more young local kids because it is Brazil and a bit different than Europe or the US (which parents let their kids go out alone on old gear?).

53

Do you mean that 36-40 years old men are more lickly to be interested by Susi Mai pictures

Will be 50 in Feb...