Sunday, February 21, 2010

Intolerance and education in Ireland

Intolerance is a terrible thing. In fact I have no time for people who are intolerant. String 'em up, I say. So I thought it would be interesting to see what are (some) of the characteristics of people who are intolerant. The instrument I use is a question in the European Social Survey which asks whether the respondent thinks "gays & lesbians should be free to live their life as they wish". Its a 5 category answer and I define a dummy which is 1 for the top 2 categories ("agree strongly" or "agree"), 0 otherwise. About 75% are tolerant by this criterion, using the Irish data. So we are a pretty tolerant bunch, according to ourselves anyway.
Below I have a simple model. The coefficients are marginal effects. So what the results show is that each additional year of education is associated with about a 7% higher probability of being tolerant (remember 7% relative to 75%). If you are reading this you are probably fairly educated so this will conform with your prior view (i.e. your prejudice): educated people are nice people. Older people are less tolerant towards gays. Well what do you expect? Likewise people who describe themselves as religious. Women are the good guys, being much more (20%) tolerant than men.
What is curious is that parental education has the opposite effect to one's own education, the more educated your father was the less tolerant you are towards gays. Parental education is a big, perhaps the biggest, determinant, of one's socio-economic status. Its an interesting paradox: answers please on a post-card...

tolerant towards gays
years education 0.0681***

age -0.00532*

woman 0.210**

very religious -0.311***

father's education -0.0823**

N 1631
Absolute t statistics in parentheses
* p<0.05,>


Gerard O'Neill said...

"Women are the good guys, being much more (20%) tolerant than men."

Of course they are - sure aren't they made of 'sugar and spice and all things nice'? ;-)

This is one of those 'zero cost' tolerances which gives the person experiencing tolerance a nice feeling of moral superiority without any obvious impact on their lives.

But when you switch to something like choosing a mate then 'tolerance' comes at a cost: hence the finding reported by Robin Hanson that women are significantly more likely than men to state a racial preference for prospective mates than men:

As Hanson would say, a lot of our 'tolerances' are simply a form of signalling. And some signals are more costly than others ...

BrendanH said...

Having an educated father is associated with many characteristics promoting tolerance (income, education). It's only the net effect is negative. What happens if you use difference between own and father's education?

Cultural discontinuity may be part of the cohort change, and this may be less from those from more highly educated parents.

Peter Carney said...

That is an interesting one.

Is it plausible that father's education is a proxy for father's religion which somewhat mediates ones religious category?

Fathers with more education may have lower religious persuasion, but, since they were educated with a Catholic ethos they are still 'somewhat' religious. Their children are thus less likely to be 'very' religious but are more likely to be 'somewhat' religious relative to the children of less educated parents who are more likely to be 'very religious'.

So, if you can tolerate the hypothesis, try including the 'somewhat' religious in the religious persuasion variable. I suspect this will lower the coefficient on parental education and solve the paradox..

do I get prize?

Alan Fernihough said...

include age^2

Mark McG said...

Manic Street Preachers can't be

Kevin Denny said...

Gerard: that doesn't explain why its higher for women then men. Even if some of the expressed difference is just talk, I would be surprised if it was all due to that. That's just a hunch.
Racial preference is not the same as intolerance: maybe you just think that blacks or whites or blondes or whatever are "hotter".

Brendan: I can't do exactly what you suggest as fathers education is in levels (not years, so I really should have had a set of dummies but it would have been messy) but if you were to somehow make the two variables comparable I am pretty sure that the positive own education effect would dominate.

Peter: doubt it, they are all Catholics pretty much. I can certainly make the religion variable more flexible (its a 1 to 10 scale). In the unlikely event that this resolves the paradox you win the prize: an autographed copy of my next paper.

Alan: tried that.Nothing doing.

Liam Delaney said...

Kevin - is there something going on with the missing values for father's education? Is the fathers education variable monotone i.e. are the levels increasing in the numbers. The sign of the coefficient looks very strange.

Liam Delaney said...

If we accept the results, the only thing I could think of is that intergenerational and cross-spouse education levels are so correlated in Ireland as to make partially them out in a regression impossible.