Tuesday, March 07, 2006
Extraordinary Claims Vs. Evidence
by Tom Bozzo
If you want to establish the proposition that the housing boom has been a relative bust for the real estate selling trade (summarized here at the Freakonomics site), you might just want to characterize:
- The relative magnitudes of entry into the profession and declines in "typical" agents' income, over the period of interest;
- Outcomes for "incumbent" agents versus entrants;
- The apparently much lower effort required to earn commissions in "boom" areas;
- The next-best alternatives for agents who manage to make some money despite what often appears to casual observers to be middling actual skill or talent (and decent money on average, neglecting the multimodality of the distribution of effort over hardcore sales sharps, physicians' spouses with time to kill, etc.).
Comments:
<< Home
If you told me that (1) there was a sudden influx of salesmen and (2) there was an increase in turnover that was less than the influx of salesmen, I would not be inclined to claim that a decline in median income reflected a decline in income.
Many of us made less money, especially inflation-adjusted from 2001 to 2004; that the median went up doesn't mean we all benefitted.
An economist who uses the median as his benchmark when the underlying dataset has shifted significantly is not one to be trusted.
Many of us made less money, especially inflation-adjusted from 2001 to 2004; that the median went up doesn't mean we all benefitted.
An economist who uses the median as his benchmark when the underlying dataset has shifted significantly is not one to be trusted.
Levitt and Dubner were annoyingly imprecise as to whether they were uniformly reporting medians versus a mix of medians and means; the paper that was in part their source (which I'm still wading my way through) is also less than crystalline.
As to your final comment, amen. That the dataset has not only shifted significantly but is also almost surely multimodal makes comparisons based on central tendency measures even less meaningful.
Post a Comment
As to your final comment, amen. That the dataset has not only shifted significantly but is also almost surely multimodal makes comparisons based on central tendency measures even less meaningful.
<< Home