Thursday, October 01, 2009

Skewed surveys

So many surveys are crippled. I've downloaded some summary results for a survey arranged by http://www.gamesindustry.com/company/542/service/1762
The results are interesting, but woefully misleading because they repeatedly quote percentage of population playing games, yet limit their "population" in two ways. The first isn't unreasonable--at least 8 years old. The second is entirely unreasonable--they only count people who have Internet access. Insofar as many forms of video games do not require Internet access, why this limitation? To make the numbers sound more impressive?

Lest you say, "everyone has Internet access", NOT EVEN CLOSE. Many many people don't even own a computer, many because they don't want to, some because they can't afford it (yes, even now when computers are so much cheaper). Some of these people may play games on phones or on friends' computers/consoles, yet why they're excluded entirely is beyond me. This also skews the comparative results of this international survey, as I'm supposing the percentage of people who have Internet access varies somewhat from country to country.

So the results are interesting for comparative purposes, but the overall percentages are mostly useless.

2 comments:

EastwoodDC said...

Internet polling is notoriously bad. Limiting it to those who have internet access may actually improve the credibility of results by filtering out some bad data.

Jolo said...

Internet polling is useless. 4Chan has skewed polls and PZ Myers (he blogs at Pharyngula) has made it a mission of his to "pharyngulate" any polls they find.

If you wanted to get a more accurate poll you would have more success getting thosed polled to either send in an email or something like that.