Tuesday, May 10, 2011

Measuring the Unmeasurable II

I do not think we are measuring what we think we are measuring...:
Gallup's Economic Confidence Index improved 10 percentage points in the week after Osama bin Laden's death, May 2-8, resulting in the best weekly confidence level since mid-February. Thirty-seven percent of Americans now say economic conditions are "getting better."
Measurement error in surveys:
Matt reminds us that there are reasons to want husbands and wives together and reasons to want them separate for questioning. Neither way is going to dominate.
Researchers sometimes forget that most of the multitude of questions respondents face in a study are extremely private. Imagine if someone from your city, claiming to be associated with some distant, foreign university, came to your door and started asking you about your detailed household expenditure, or your spouse about their sexual history?
Measuring gender fairness the Dilbert way (ie - Scott Adams said this) (HT: Newmark).
So I propose a simple test to determine if you, individually, are a victim of gender unfairness. If a genie gave you the chance to magically switch your gender, and become a member of the other sex, would you do it? And let's say the new you would be about the same as now on the scale of attractiveness, intelligence, ethnicity, circumstance, and health. The only real change would be gender. Do you take the offer?
If your answer is no, then maybe fairness isn't what you really want. Maybe what you want is all the advantages you have now plus the good stuff that other people have. I totally understand. I want the same thing.

No comments:

Post a Comment