(I hope that says Unfortunately I can't read Finnish) Hence it is difficult to reply to the Finnish parts of Mikko Ellilä's post where he picks holes in my post picking holes in the post of his that got him in hot water with the multicultis in the Finnish government. This post covers other blogospheric reaction and I am happy to say (despite what this may imply to those who like tarring people as racists or sexists) that I actually agree with a good deal of it, in particular his nailing of the sort of strawman argument that goes "if you think all blacks are dumb criminals then how about Condoleezza Rice (or ...)". I'll go further and try to explain the concept of normal distribution, probability, statistics etc. because this is something that your average arts graduate fails to grok.
If I state that "on average population X is more Y than population Z is" then this is not a statement that all members of population X are better at Y than all members of population Z. And this is the case no matter what X, Y and Z are. It applies when X and Z are Europeans and Africans respectively and Y is intelligence just as much as it applies where X and Z are smokers and non-smokers and Y is dying young. There are folks who smoke a packet a day and live until they are 101 but if you look at the population of centenarians you will find that far fewer of them smoke (or have smoked for a long time) than the general population. Condoleezza Rice is a) black and b) female but she is smarter than a lot of white males. But the question should be whether at her level of smartness (measured in some way) are there more whites and males that blacks and females. In other words if, in the total population, there are an average of 2 white males for every 1 black female (numbers taken for example not based on actual data) do we find the same ratio when it comes to people of her smartness? or do we find that the ratio is now 10:1 instead of 2:1? Failure to grasp this point led to Larry Summer's resigning (he asked approximately that question his critics failed to understand what he was asking) and it seems to be difficult to grasp but it is critical none the less. Ann Althouse has written a number of times (e.g. here) that the scandal is always in the presentation. Saying that "%oppressed minority% is actually better as %activity% that whites/men/%some unoppressed group%" is alwyas going to be OK whereas saying that they are worse is a no-no.
Indeed as I wrote above, other than the bit referring to my post I actually find myself agreeing with Mr Ellilä. The part where we seem to be at some sort of disagreement is over Mr Ellilä's tying of race/religion/immigrant status to criminality. I'm not denying that the statistics show that in Finland the immigrant population is more cimminal than the native one or that the imigrant population is of different ethnic origin to the native one what I am complaining about it Mr Ellilä's apparent confusion of statistical correlation with causation. He may turn out to be right (I don't know) but I don't feel that the evidence he presented was convincing and I don't feel that his response is either. In order to find out whether it is ethnicity, religion or factor X which leads immigrants to commit more crimes you need to control for these factors and look for examples where other groups of differing religion/ethnicity/factor X have immigrated and see whether they are as prone to criminality. Mr Ellilä seems to understand statistics quite well so I'm sad that he seems to miss this point.
Blog Home : June 2007 : Permalink