?

Log in

No account? Create an account

Previous Entry | Next Entry

A kind person has pointed me (via a locked entry in her own lj) to the report discussed in my previous entry. It appears that I misread the key piece of information:
For both the [abstinence] program and control group youth, the reported mean age at first intercourse was identical, 14.9 years. This age is seemingly young, but recall that the outcome is defined only for youth who reported having had sex and the average age of the evaluation sample was less than 17.
Getting means and medians sorted out: The key statistic in most of the countries I quoted in my previous entry is (as I should have guessed) the median date for first sexual intercourse, since it is established once the 51st percentile have lost their virginity, no matter how long it takes the other 49% of your sample.

In the case of the American survey, 49% of the sample were still virgins at the end, and we are told that the mean and modal age of all surveyed was 16, so I guess that will make the median date for first sexual intercourse of the whole sample also around 16; which is low-ish by international standards, but not quite as unusually low as 14.9.

Comments

( 4 comments — Leave a comment )
bennmorland
Apr. 17th, 2007 07:08 am (UTC)
I am not surprised by my lack of surprise before, at the 14.9.

Does this make sense?
watervole
Apr. 17th, 2007 07:37 am (UTC)
I'll have to remember that. I've been seeking good examples for my math pupils to explain why a median may sometimes be more useful than a mean.
artw
Apr. 17th, 2007 09:28 am (UTC)
My son's school test results routinely come back with the class median score alongside his own score. While baby growth charts are usually marked out in percentiles, not standard deviations from the mean, his school marks show the advantage of the median better I think: easy to calculate for a small sample, and not distorted by outlier scores. But statistics about sexual intercourse will surely hold your pupils' attention better. Did you see University Challenge last night, the expression of smug disbelief on the face of the maths student asked to name the Greek letter used in statistics to indicate a sum?
watervole
Apr. 17th, 2007 09:42 am (UTC)
More to the point in this case, it would be impossible to calculate a mean until all the children had lost their virginity. As soon as half of them have, you can calculate a median. It's an advantage I'd never previously though of and it can be applied to a whole host of things that involve life-expectancy, etc.
( 4 comments — Leave a comment )

Latest Month

March 2019
S M T W T F S
     12
3456789
10111213141516
17181920212223
24252627282930
31      

Tags

Powered by LiveJournal.com
Designed by yoksel