Tuesday, August 01, 2006

Go Get 'Em, Lefty!

A new paper by Christopher S. Ruebeck, Joseph E. Harrington, Jr. and Robert Moffitt studies the job-market outcomes of left-handed vs. right-handed people:
We examine whether handedness is related to performance in the labor market and, in particular, earnings. We find a significant wage effect for left-handed men with high levels of education. This positive wage effect is strongest among those who have lower than average earnings relative to those of similar high education. This effect is not found among women.
A strange result... why would this effect exist? And why would it exist only for men, and not for women? And for highly educated men, but not for less educated men? I wonder if this is not just a "random" result.

Unless someone has a reasonable theory, I'm likely to dismiss the result as a strange fact, likely due to randomness in the data (the relationship between education and handedness has a t-stat of only 2 in the first regression, Table 2, and more detailed looks at this result that follow also seem to have just-signigificant statistics). For what it's worth, the authors also fail to provide a reasonable theory:
We explore some possible explanations for these findings but are not able to provide concrete evidence leading to a theory that can reconcile all of the various facts we identify. We recommend this as an avenue for future research.


Blogger BW said...

I've long wondered about this. I've even tried to look at it in some data, but clearly I didn't push hard enough. I first started to wonder about this during the first year of grad school. Nearly 1/3 of my grad school cohort was left-handed, and (at least among those I noticed) there are way more left-handed faculty in the Harvard economics faculty than the 1 out of 9 there should be.

8/01/2006 5:58 PM  
Blogger BW said...

Now that I've read the paper, It really is a weird collection of results, but I think there explanations aren't totally crazy. Ideally, someone would add the handedness question to the CPS for a few months so we could get both more observations and more detailled occupation info, etc.

8/01/2006 7:11 PM  
Anonymous Anonymous said...

Theories? I love theories!

So here's one;

a) Correlation is rarely causation, especially in economics :).

b) Handedness also correlates well with Brain...ness? See Left Brain/Right Brain cognitive theories, or a book like:

c) There is a distinct difference between the utility of the cognitive output from 'right brain' and the utility of the cognitive output of a 'left brain'.

d) There is a distinct difference between the value (demand curve) of the cognitive output from 'right brain' and the value (demand curve) of the cognitive output of a 'left brain'.

e) The value of 'right brain(left handedness)' cognitive processes is increasing faster than the value of 'left brain (right handedness)' cognitive processes.

f) This is evidenced as a positive wage effect for Men who use their left hand to write.

I think 'higher education' would support this theory, as higher education correlates well with evolving job markets, i.e. new industries, new job functions, new opportunities, etc.

The fact that it didn't have the same effect for Women suggests that this theory is a) totally in left field, or b) that there is a factor that causes the effect to show up in one gender and not the other.

There is a path here that I won’t go down, but I think we would need to understand all the factors that go into the current wage disparity between Men and Women before we would be able to explore all the different factors that affect a person's wage, and yes, there is a fair amount of studies that say these factors are different for the two genders, and if the groups of factors are different, than its totally reasonable to believe that an added factor could have different results when mixed in with these two distinct groups of factors.

If that makes any sense.

But more importantly, correlation is rarely causation, but in a similar way, the lack of causation is rarely enough to claim a lack of correlation.


8/02/2006 5:03 PM  

Post a Comment

<< Home