Tuesday, October 11, 2016

I miss the old NPR

For years, decades even, NPR was a staple of my daily routine. Fresh reporting, not just repetition of the same stories everyone else was carrying. Variety. Deep research. Some really good stuff. You couldn't listen for half an hour without learning at least something and usually a lot.

But over the years the quality has declined. They have always leaned left, or at least have always had the mind-set and the unquestioned assumptions of wealthy, upper-middle class, highly educated, urban dwellers. They did a reasonable job trying to consciously balance that a bit, long ago but in recent years, there has been less and less effort.

More to the point, their reporters seem less and less intelligent. They seem happy to start with improbable assumptions and then speculate crazily from there.

I came across Researchers Find Racial Wage Gap Has Grown today which illustrates the decline. Its just SJW press release journalism, no skeptical mind required. The original report is Black-white wage gaps expand with rising wage inequality by Valerie Wilson and William M. Rodgers III.

For years, feminists have been trying to make the claim that women are paid less for the same work as men. Claims that have been systematically debunked by economists as well as academic feminists (see the work of Claudia Goldin for example.) The claim rests on rank innumeracy and comparing apples to oranges. If you want to see whether women are being discriminated against in terms of compensation, you have to compare like to like. Same education attainment, same number of hours worked, same number of years worked, same field of endeavor, same area of study, same industry, same number of career interruptions, etc. When you compare like-to-like, there is no evidence of wage discrimination.

This has been demonstrated again and again. There is no reason for someone in media not to know this. You have to control the variables to get like-to-like.

And what do we have here? The same SJW agenda-driven sloppy research that compares apples and oranges to come up with the conclusion they had already reached; African Americans must be being discriminated against because they earn less overall.

The researchers seem to be aware of the perils of their flawed approach. They try and make the claim that they controlled for the important variables:
A big part of what we did in this study was to try to pin down all of the factors that would reasonably affect what workers get paid. We controlled for level of education. We looked at years of experience, what region of the country you live in, whether or not you live in an urban area or a rural area. And even after controlling for those factors, there was a sizable difference between what black and white workers were paid.
Well, fine. But those aren't the most important variables that determine outcomes. Number of hours worked, academic field, industry, sector, continuity of employment, etc. Control for those and you have something meaty to report on.

Take, as an example, one of their four controls - education attainment. Doctorates in Education, Psychology and Sociology are a dime a dozen and among the least paid fields. Doctorates in Computer Science, Artificial Intelligence, and Electrical Engineering are among the highest paid fields. If any group is overrepresented in the former three and underrepresented in the latter three, they will have dramatically lower average compensation for the same level of education attainment. By a factor of 2 or 3. And indeed, African Americans are overrepresented among the soft sciences and significantly underrepresented in the hard sciences. The difference in compensation has nothing to do with discrimination and everything to do with the fields they chose to specialize in.

Since the researchers seem to recognize the importance of controlling determining variables but elected not to control for the important variables, I am inclined to believe that they were after a result consistent with their predetermined SJW biases.

But that's just SJW academics being SJW academics. Nothing too out of the ordinary there, even if undesirable.

What is outrageous is for NPR, knowing that these type of studies routinely omit the important variables to reach predetermined conclusions, elected to broadcast the results anyway and without pressing the researchers on why they chose to omit the important variables.

Shame on NPR. Increasingly just a conveyor of cognitive pollution. I really hope they right the ship and start coming back to hard hitting, fact-based reporting in the future. I miss the old NPR.

No comments:

Post a Comment