Field of Science

How good is the evidence?

I am a member of the Brent Teenage Pregnancy Partnership Board I keep my eyes open for articles that might have some relevance to the task of trying to reduce the number of teenage pregnancies. There is clearly also a connection with preventing STDs (sexually transmitted diseases) .

Today I came across this article with the headline:

"Quarter of U.S. teen girls have sex-related disease"

at the end on the 2nd page I read:


"The findings were based on data from 838 girls who took part in a nationally representative health survey in 2003 and 2004. They were tested for various STDs."

I have never studied statistical analysis, or really understood the principles of carrying out sound surveys etc. but apart from noting the fact that the survey is already 4 years out of date I have a gut feeling that 838 girls isn't a large enough sample.
Any comments?
p.s. I have made a small effort to find the original report but without success

1 comment:

  1. Josh, the study hasn't been published but the press release is here. Doesn't give any extra info though. 838 is a pretty good number - opinion polls typically have only 1000, and have a margin of error of a few percent. The biggest problem with this one is recruitment bias - i.e. it's only those who took part in the survey. Those who didn't want to don't appear in the numbers.

    ReplyDelete

Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="http://www.fieldofscience.com/">FoS</a> = FoS