Friday, December 17, 2004
Dumping large amounts of hormones into your body on a daily basis for years at a time doesn't reduce risk of heart disease
Researches befuddled. Incidentally, I can maybe explain whlie the old study found a correlation where this one didn't. Check this out.
That’s because it relies on women’s memories of what drugs they used in previous years rather than actual hospital or medical records.
As I learned in Econometrics, when there's measurement error, no matter which way it goes, all estimators of the effect of independant variables on your dependant variable become biased upwards. There is no way to correct for this, either, except getting better data or a proxy variable to replace the errored data. Extraordinarily counterintuitive, but the math works out, and it happens in reality all the time.
That’s because it relies on women’s memories of what drugs they used in previous years rather than actual hospital or medical records.
As I learned in Econometrics, when there's measurement error, no matter which way it goes, all estimators of the effect of independant variables on your dependant variable become biased upwards. There is no way to correct for this, either, except getting better data or a proxy variable to replace the errored data. Extraordinarily counterintuitive, but the math works out, and it happens in reality all the time.
