shouting loudly

Survey: Illegal downloading common, not seen as serious

In a nationally representative sample conducted last year, Solutions Research Group found that millions of Americans are downloading movies and music illegally, and a majority think it is not a serious offense.

The study found that 32 million Americans age 12 and over have downloaded at least one full-length movie. The typical movie downloader is 29, male (63/37 split), has 16 downloaded full-length pics on their PC, and gets all of their downloaded movies illegally from P2P networks.

The attitudes toward downloading are also fascinating. About 2 in 5 thinks downloading music or movies illegally is a “serious offense,” but 4 in 5 feel that way about pocketing a DVD from a store. People are more likely to be concerned about parking in fire lanes than illegal downloading.

So much for the movie industry being ahead of the internet curve. It was a mere bandwidth issue that kept their material offline when Napster had Lars Ulrich begging customers to pay, but some people (including many in the biz) thought the extra time would give the studios a chance to get together a proper business model.

Turns out, the major movie studios are in the same position as the record labels: pushing iTunes downloads, fighting a legal and PR war, and wondering if they could have done anything differently. (Hint: YES!) Meanwhile, the internet has unleashed an unstoppable torrent of indy music and movies, opening up creative and financial possibilities for those without access to much capital.

BONUS: Triple props to SRG for this survey. First, good combination of questions on a great and very current topic.

Second, good step toward Creative Commons-like permission to use their graphic (”Note: May Not Be Reproduced Without Proper Attribution”). This isn’t the most free license possible (it would be nice to make the permission part explicit), but it ain’t bad.

Third, and most obscurely, check out their description of the margin of error. “The results cited in this release are accurate to plus or minus 2.4 percent, 19 times out of 20.” They are 95% certain that the results are within that margin.

In other words, if they did 100 random surveys with the same questions, the answers they got would be within 2.4 percentage points of the first survey 95 times. (The other 5 would probably be close, too, but it’s statistically possible for your copyright poll to randomly dial 1000 recording artists and movie studio executives, so there’s no guarantee.)

This is why you should never listen to people who suggest that a 2% change in a regular survey item–say, the presidential approval rating–necessarily matters much at all. A 2-3% change in a survey item is just as likely to be statistical noise. Even if there is a change that’s outside the margin of error–let’s say a 4% change in a measure with a 3% margin of error–we’re only pretty sure that it’s a real change and not statistical noise. 19 times out of 20, it would be a real change, but as much as 5% of the time, a change can be “statistically significant” and actually still be statistical noise.

If you have ever written or will need to describe the results of a survey but you’ve never taken statistics, I recommend you read more about margin of error.