31% is statistically significant (roughly 3 sigma) for a base rate of 100--sqrt(N) is the expected variation in a random sample with mean value N--so called "shot noise." But as you say, it's only marginally significant--right at P-value 0.05. That said, the increase for the rates against Muslims in NYC is easily significant. The increase for Jews is not (10% of 100 is 1 sigma--not remotely significant, even for a single-tailed distribution, which is arguably valid when the original hypothesis was of a positive change.)
I suppose statistics is a requirement for all social sciences at your school, including philosophy? I know logic is a requirement, but it seems to me that statistics and game theory are pretty much a necessity today.
Yeah, I know philosophy isn't exactly a social science, though often it's impossible to distinguish from economics. (Nozick, Hayek, Mills, Rawls...) But any analysis of actual data requires statistics. And game theory often illuminates the line where econ ends and phil. begins.
31% is statistically significant (roughly 3 sigma) for a base rate of 100--sqrt(N) is the expected variation in a random sample with mean value N--so called "shot noise." But as you say, it's only marginally significant--right at P-value 0.05. That said, the increase for the rates against Muslims in NYC is easily significant. The increase for Jews is not (10% of 100 is 1 sigma--not remotely significant, even for a single-tailed distribution, which is arguably valid when the original hypothesis was of a positive change.)
ReplyDeleteI suppose statistics is a requirement for all social sciences at your school, including philosophy? I know logic is a requirement, but it seems to me that statistics and game theory are pretty much a necessity today.
"I suppose statistics is a requirement for all social sciences at your school, including philosophy?"
ReplyDeleteA strange question, in several ways...
Yeah, I know philosophy isn't exactly a social science, though often it's impossible to distinguish from economics. (Nozick, Hayek, Mills, Rawls...) But any analysis of actual data requires statistics. And game theory often illuminates the line where econ ends and phil. begins.
ReplyDelete