We actually agree with a large swath of IRB bad arguments, which is what this little spat is about. But still:

The idea that the "typical journalistic interview or survey" involves great questions of war and peace and commerce and culture, while the typical social-science interview or survey involves dog owners, is, I suppose, a testable hypothesis, though not one that Weinstein tests by sampling news stories and journal articles. Absent such evidence, it is an anti-intellectual slur.

There's an extent to which the ethos of some social scientific scholars - with their occasionally fetishistic insistence on methodological rigor and bureaucratic niceties - deserve an IRB process. But probably not the IRB process that actually exists, since nobody deserves that. See, inter alia: here and here and here.

References:
* James Weinstein's Anti-Intellectualism [IRB Blog]
* Study Finds IRBs Exaggerate Risks of Survey Questions [IRB Blog]
* Roberta S. Gold, “None of Anybody’s Goddamned Business”? [IRB Blog]
* Guidance Creep [IRB Blog]

Previously:
* Quantitative Hints Of The Big Other
* Disturbingly, Quantitative Analysis May Have Methodological Shortcomings - Medical Studies Edition
* Disturbingly, Quantitative Analysis May Have Methodological Shortcomings - Blogosphere Edition