Study of the Net's impact on music sales

Edward Felten describes, and links to, a Princeton student's senior thesis which studies the claimed "filesharing is killing the record industry" effect from another angle:  because it's so difficult to get accurate data on people's filesharing habits (just like sexual behaviour:  they have too many reasons to lie on the surveys, or to incorrectly assess their own behaviour even if they're trying to answer honestly), Eric Boorstein studied the effect of Net access in general, not filesharing in particular, on sales of music.  Controlling for income level and several other factors, he found that Net use decreased music purchasing among young people, but increase it among older people by enough that the overall effect across the entire population was an increase in music sales correlated with Net use.  Both the original study and Felten's analysis go into some detail on just what that means, and the ways in which the conclusion must be treated with caution.  [Study of the Net's impact on music sales]