>"Data on wages, for instance, cannot be sampled
>randomly if only individuals with certain characteristics --
>unobservable to the researcher -- choose to work or engage in
>education,"
Er, are they saying that if your sample isn't random your results may be skewed?
> ... estimating how the duration of unemployment affects
>chances of getting a job. Heckman is also a leader of applied
>research in these areas."
Let me guess ... the chances diminish, right?
>"Data regarding individuals' occupation or place of residence reflect
>choices they have made among a limited number of alternatives,"
No shit!
>"Prior to McFadden's contributions, empirical studies of such choices
>lacked a >foundation in economic theory.
Never mind the foundation it has in the bleedin' obvious ...
>Evolving from a new theory of discrete choice, the
>statistical methods developed by McFadden have transformed empirical
>research."
... so empirical research is finally getting to the stage where it can come to agreement with the bleedin' obvious, eh? Well, we shouldn't mock. It's found the otherwise plenty of times before.
>Examples of McFadden's extensive
>applications of his own methods include the design of the San
>Francisco BART [Bay Area Rapid Transit] system,
Lemme guess again ... more freeways because more cars were expected, going to and from places people seemed to want to go?
>as well as investments in phone service and housing for the elderly.
I hope he found that these would be a good idea ...
Wish they could tell me why I keep sending babble to LBO when I've other things to do ...
Cheers, Rob.