09 November 2010

Tiny Little Cocktail Link Sausages

Just a few cocktail weenies this morning, as I have remora duties...

  • Where does W's memoir belong? Probably not in "serious nonfiction"; I nominate "religious and inspirational," myself...
  • For those of us old enough to remember VHS v. Beta (and I do not mean the court battles, but only the marketplace), figuring out which incompatible device and incompatible software standard to support in e-readers is better left for another time. particularly for those with wide reading interests or — at least until now — read extensively from periodicals and care about proper payment.

    For me, though, it's not just my inherent cheapness; it's battery life. I should not need to take yet another cable (or dedicated charger) with me on a one-week business trip or vacation and rely on having good power at the other end... and in transit at airports, on planes/busses/trains, etc. When these devices have ten-day standby with at least 100 hours active in that standby time, I'll start considering an e-reader... that also reads standard third-party formats (PDF, TXT, RTF, DOC, HTM) without requiring me to use a proprietary "store" for material that I already have and without reporting to the manufacturer/sponsor what is on the device or allowing any third party to delete or add content. (EPUB is not, contrary to its advocates, a "standard" format — at least not yet, thanks to inconsistent handling of graphics and notes.)

  • The Market Design Blog has a fascinating piece on market incentives for erroneous medical research that also has some disturbing implications for so-called "empirical legal studies" — you know, those "surveys" of court-case results that purport to tell us something about litigant motivation. Leaving aside for the moment that these are epidemiological, not empirical, studies (and anyone who has ever studied stochastics or public-health statistics knows why that's important), it is truly disturbing how many of the same biases and analytical weaknesses Dr Ioannidis describes in medical research appear in "empirical" legal (and, so far as I can tell, economic and financial) research studies... especially, one might add, in post hoc analyses of market failures. As Professor Roth emphasizes in his post, though, this is an inevitable consequence of the scientific method. The problem is not with science per se; it is with the incentives for exploiting scientific results in the marketplace (whether the financial marketplace or the reputational marketplace is irrelevant).

    What I find most interesting, though, is the implications that Dr Ionnadis's work carries for the entertainment industry, and in particular for the publishing industry. As a specific example, consider whitewashing of covers. As I have noted, I'm satisfied that the source of the whitewashing phenomenon is misinterpretation of a chain buyer's offhand comment. Dr Ionnadis's work, though, points at a much more disturbing, second-order problem of causation: What data did that chain buyer use, and what was that chain buyer's analytic method, in coming to that conclusion? Similarly, what data do S&M dorks use, and what analytic process do they follow, in coming up with cover memes like "metallic-ink lettering sells more books," or the stereotypical H'wood "arrange the cast in a V around the highest-paid actor" publicity photo... especially when it is slapped on the cover of a periodical with a feature article about that H'wood property or a novelization? In short, is this bad data, bad analysis... or no analysis at all? I tend toward believing the last, particularly based upon the source of much of the received wisdom in sales (based on immensely flawed, nonreplicable, nonverifiable studies of fungible commodities being misapplied to nonfungible items and services being sold to a different population).