Links of Notice
Michael on the Heritage Foundation’s crap calculations: publishing papers with statistically non-significant results, and then telling people to “decide for themselves.” A new low, the stats community should be outraged.
or
Michael on the Heritage Foundation’s crap calculations: publishing papers with statistically non-significant results, and then telling people to “decide for themselves.” A new low, the stats community should be outraged.
This is a medical weblog--a collection of thoughts about medicine, medical training, and health policy--written by a fifth-year medical student.
I recently stopped blogging, as I graduated from medical school and I'm now a physician in my residency training in New York City. But feel free to read and enjoy!
(To get rid of this thing, just wave your mouse over it. Ta da!
That’s ridiculous! Besides the obvious (involving simple common sense), anyone who’s taken even an elementary stats class knows a stat sig at .10 is far too high for any testing. Ugh.. me raging. How many thousands of kids going through these programs with the only information that have concerning sex if from their equally ignorant peers?! Its not only insulting and dangerous, but flat out irresponsible! Going beyond this, did you hear about Wisconson try to ban the morning-after pill on state college campuses? (http://apnews.myway.com/article/20050617/D8APELD00.html) Of all places (especially since the college years have a far higher incidence of date rape). So angry.
Uhh no. The results ARE statistically significant, just at a lower level of confidence. Plenty of acceptable social science research is published with those levels of significance. I think you’re outrage is more based on politics, than science. After all, non-abstinance sex education has not been shown to reduce STD s or risky sexual behavior.
With a sample size of 15,000, that’s plenty of power to get a more acceptable p value. I don’t buy it.