« More Quake papercraft | Main | UK programmer WLTM UK developer »

May 20, 2006



I was intending to spend less time gaming in 2006 than in 2005. Then I got WoW.

This is the fundamental problem with intentional studies.


Yeah, Alice - I found a similar statistic when researching my ETech talk, but I'm not sure _where_ I found it. I'll go and dig it out, but it was very similar numbers-wise, and based on very similar questions.

It also makes statistics about general console-uptake, etc, seem even more interesting, when this supposedly "core market" is quite possibly on the wane.

700 people isn't a sample size for ANYTHING that purports to be about what "teens" are doing. There's several million of them, you know?


One of these days I'm going to remember to type in all my info BEFORE hitting submit. *sigh*


Intentions are one thing but actions are another.

This is the worst kind of survey: do we even know what the situation was for the interviewees? Was the interview conducted by the kind of person who the teens might have thought disapproved of gaming? Did they feel compelled give that answer? In all honesty I'd probably say that I intend to play less and read more every year, but Christ knows whether I do or not. Does anyone actually keep an accurate record of how much time they spend doing something?


Put that Pr0n in games and you'll get the teenagers back lol

I think one of the problems in the US, is that they keep re-rating games up to the hilt for one retarded reason or another, and now there are only 6 titles in the US than a Teenager could purchase!

100% Fact!

Also, 700 is a good sample size. I sampled 700 Smarties and concluded they all taste the same regardless of colour. It's totally the same thing, don't look at me like that.

100% Fact!

I love surveys and statistics. They actualy retard the progress of mankind.

Robert 'Groby' Blum

Sigh. 700 might very well be a good sample size. It depends on the error, confidence level and variability. If we have maximum variability (i.e. we can't predict the answer, there's no clustering), a 5% error (i.e, in reality it might be 75%-85% of teens in the overall population) and 95% confidence level(i.e. there's a 5% chance this is wrong), a good sample size for large populations is 385 - a bit more than half of those 700.

Now, the sample selection process plays a huge role in the outcome, too - not having spend $30, I've got no idea what the process was.

But in general, sample sizes for accurate answers are surprisingly small. So next time, before you dismiss a sample size as "very small", you might want to ask somebody with a background in statistics.

That's not me, btw. All I recall is a couple of semesters at Uni, but it's not my profession.. So take my numbers with a grain of salt. But *do* ask somebody who knows the actual science.

The comments to this entry are closed.

Recent links