Hacker News new | ask | show | jobs
by antiterra 4861 days ago
That kind of statistical hand waving is irritating. First, it's 40 abductions where parents are characterized as being paranoid about their child being kidnapped. The statistic says nothing about how leaving your kid on a sidewalk might affect the chances of abduction. Your chances of getting hit by lightning aren't as slim if you're a golfer.

Also, 700x more likely to get into Harvard as an applicant or just a US born child? Does that include post-graduate Harvard institutions? How is 'getting into Harvard' a useful benchmark for acceptable risk?
1 comments
rz2k 4861 days ago
Clearly leaving a child unattended is going to have an effect, and so are many other factors going to have an effect on admission to Harvard.

Furthermore, the hand waving statistics are off by an order of magnitude as far as I can tell. 40 * 700 = 28,000, but the cohort of births in a given year result in a student body closer to 2,800 than 28,000.

link
antiterra 4861 days ago
It probably came from comparing accept percentage of Harvard (6% now but 7% a couple years ago) with a kidnap chance calculated for the whole of children born in the US. Of course applicants are already a selected subset and not representative of all candidates born in the same year. It's a pretty useless comparison.

e.g. 40 kidnaps per year for 10 years = 400. 400 * 700 'times more likely' = 280,000. Divide that by 4mil children born in a year to get 7%.

link