Statistical Bias in MacHeist Participants
Sunday, 2007-12-02; 17:48:00
While practically all to be said about MacHeist has been said between me, Guy in the Hat, and the comments, I would like to emphasize one important point.
In his third entry about MacHeist, Guy in the Hat has this to say:
There's an important statistical error being made here, because there's a strong positive correlation between those developers who participated in MacHeist and whether they support MacHeist or not. This isn't a coincidence — if a developer had a problem with MacHeist's terms or didn't see any value in it, they wouldn't have participated. So practically by default, any developer who participates in MacHeist will support MacHeist.
Guy in the Hat's tone here seems to imply that because Mueller, Gruber, and Kafasis didn't participate in the promotion, their opinions matter less than those of the developers who did participate. But because of the strong correlation between those who participate and their opinions on MacHeist, this basically belittles the opinion of anybody who doesn't like MacHeist: 100% of the people who don't like MacHeist didn't participate or aren't a developer.
Don't get me wrong, I'm not dismissing the arguments of those who participated, either. Obviously there was a strong incentive for those 10 developers to participate in MacHeist, and I would tend to agree. If I as a developer (and, don't forget, I am a developer) were approached by MacHeist and asked to participate, I would be strongly inclined to do so. My products are not widely known at all, so the publicity offered by MacHeist would put me on the map as a small developer. That alone is worth a lot more than the flat fee paid to developers.
But that's not where I have a problem with MacHeist. As a consumer I have a problem, because I don't like where my money is going.
In his third entry about MacHeist, Guy in the Hat has this to say:
I think most people approach this issue as “would I personally make a deal with MacHest in the same situation?” But that will never matter unless you actually did make a deal. The only people who criticized MacHeist (Gus Mueller, John Gruber, Paul Kafasis, et al.) were those who didn’t participate.
There's an important statistical error being made here, because there's a strong positive correlation between those developers who participated in MacHeist and whether they support MacHeist or not. This isn't a coincidence — if a developer had a problem with MacHeist's terms or didn't see any value in it, they wouldn't have participated. So practically by default, any developer who participates in MacHeist will support MacHeist.
Guy in the Hat's tone here seems to imply that because Mueller, Gruber, and Kafasis didn't participate in the promotion, their opinions matter less than those of the developers who did participate. But because of the strong correlation between those who participate and their opinions on MacHeist, this basically belittles the opinion of anybody who doesn't like MacHeist: 100% of the people who don't like MacHeist didn't participate or aren't a developer.
Don't get me wrong, I'm not dismissing the arguments of those who participated, either. Obviously there was a strong incentive for those 10 developers to participate in MacHeist, and I would tend to agree. If I as a developer (and, don't forget, I am a developer) were approached by MacHeist and asked to participate, I would be strongly inclined to do so. My products are not widely known at all, so the publicity offered by MacHeist would put me on the map as a small developer. That alone is worth a lot more than the flat fee paid to developers.
But that's not where I have a problem with MacHeist. As a consumer I have a problem, because I don't like where my money is going.
Technological Supernova Rants Older Newer Post a Comment