What science should not look like

[simgiran]

About Gene Abel study (which is used as source for work us as source at en wiki)
http://www.cjat.org/ipb/index.php?showtopic=926&s=6f89bc241c7035430cf8f0e5098a5ca6

[Gabriel Svoboda]

Very interesting article. I almost fainted when I had learned some people seriously think a typical paedophile commits at least 100 child sexual abuses during his lifetime. Now I know where this number comes from, and how flawed it is:

1) The sample is biased: it consists of people who de jure are not criminals, yet they are quite close to being criminal, since their behaviour has already caused alarm to law/health/social workers.
2) "Average paedophile" is not the same thing as "median paedophile". If there are 9 paedophiles who have never abused any child and a phantom paedophile who has commited 1000 abuses (which does not equal to "who has abused 1000 children"), the average will be 100. That's not fiction, the sample really contained such phantom paedophiles.

[Jednorozec]

For me nearly imposible thing. Abuse 1000 people?? Looks like willfull missrepresentation for me. Plus if i do some statistic research, i always remove such extreme results (which is normal)...