Sunday, January 25, 2009

Immigration drives some to suicide: Here's a cross-national study abstract that reports a .13% increase in suicide for each 1% increase in immigration (supposedly due to lower social integration and culture shock).


Jim Bowery said...

That's interesting and it is even more interesting that I just ran a similar ecological correlation suite on the United States, I got the following top Pearson correlation coefficients:

89 with -PopulationMultiple2000to2001
65 with -ForeignBornMultiple1990to2000
61 with -TotalFertilityRate2003
61 with -FertilityRate2003
59 with -ImmigrantsMexicoPercapita1998
58 with -MalesPer100Females2000
56 with -AdultMalesPer100AdultFemales2000
55 with -BirthRate2003
44 with -AgeUnder18Percentage2000
43 with MedianAge2000
43 with -MethamphetamineUsingPercentHighSchoolFemales2001
42 with Age65AndOverPercentage2000
42 with -AirTrafficNonScheduledPassengersPercapita1999
41 with -WhiteTotalFertilityRate2002

Now, I'm probably one of the most anti-immigration people you'll ever meet, but it's hard to imagine a more contradictory result!

What's going on here?

Jim Bowery said...

Of some interest here is that I just ran -ForeignBornPercapitaMultiple1990to2000 and the r went from 65 down to 40 -- not as strongly contradictory but an anti-correlation.

Anonymous said...

Could you enlighten us statistically challenged on what a Pearson correlation coefficient means?

Jim Bowery said...

Pearson correlation coefficient is what people usually mean when they say "correlation". It's basically how well two variables (x,y) form a perfect line and whether that line has a positive or negative slope. It's basically how we know anything about the world (induction), and we know nothing of the world by pure implication.