On Tue, 05 Aug 2014 08:27:27 -0700, Rick Froman wrote:
We now have a new happiness equation
http://www.pnas.org/content/early/2014/07/31/1407535111.full.pdf+html
to replace the discredited positivity ratio:
http://chronicle.com/blogs/percolator/the-magic-ratio-that-wasnt/33279
A few points:
(1) Any research that involves complex mathematics for analysis and/or
modeling should be explainable in plain language though it may take
some effort to characterize it properly in words. But just relying on
the
words and not checking the math is a recipe for disaster -- the
positivity
ratio is a case in point. I am not surprised that Daniel Gilbert, the
Prudential
commercial shill, was taken in nor Seligman. We have seen this movie
before, in fact, the movie is "Margin Call", about a Wall Street
investment
bank that is about to be wiped out if it doesn't sell all of its
mortgage backed
securities -- no one was aware of this until one of the firm's analysts
(played
by Stanley Tucci) came to that realization but he is "separated from
service"
because of a "downsizing" and is not allowed to warn anyone except for
a
colleague who he gives a USB drive containing the analysis as he is
escorted
out of the building. The younger colleague (played by Zachary Quinto,
the new
Mr. Spock in the Star Trek reboot) who is highly Math competent (he
actually
worked for NASA prior to joining the bank; banks pay better) completes
the analysis and modeling and comes to the realization that the mortgage
backed securities the firm has are so leveraged that if the assets
decrease
by 25% (i.e., mortgages are foreclosed), the losses will be greater than
what
the investment bank is worth (which happened in real life). The key
scenes
are where higher levels of management are informed of the situation:
when
first shown the mathematical model and results they typically say "I
can't follow
that, put it into words", all the way to the top guy in the bank. Once
the math
is confirmed, the bank does the only thing it can do to survive: they
sell all of
the securities by the end of the next day. They're still in business
though they have
screwed a large number of people with the junk securities they sold
them.
See the Wikipedia entry on the film:
http://en.wikipedia.org/wiki/Margin_Call_%28film%29
See the movie. By the way, I hear that Gilbert, Seligman, etc., are all
still in business too.
(2) I've only skimmed the Rutledge et al PNAS article but if I am not
mistaken (I may very well be), the equation they provide for "Happiness"
on the second page was constructed to take into account the effect
of prior trials on the current trial (one of the problems with
psychophysics
historically is that this has not been controlled for which makes
Fechner's
Law, Steven's Law, and similar constructions at best approximations).
But since Rutledge et al seem to argue that prior trials do not have an
effect on a current trial, this equation should simplify in form because
it should only model the current trial (if t represents the current
trial and j
represents prior trials, there is no need for t-j terms). But Rutledge
et al
provide no such simplified equation. Perhaps I will have to do a closer
reading of the article.
(3) Don't you hate it when researchers use software for analysis that
you don't have and, more importantly, is costly to get? Back in the
1950s some researchers claimed that simulations of cognitive processing
that Herb Simon and Allen Newell reported (e.g., the Logic Theorist)
were frauds because they did not believe computers could do such
analyses. Of course, Simon and Newell was willing to send the punched
cards for their programs but this was pretty useless if one did not have
the requisite IBM mainframe to run it on. Of course, researchers with
such
resources did manage to replicate Simon and Newell's work (I think).
Anyway, Rutledge et al. use MATLAB for the analyses as well as having
"tested a wide variety of alternative models". One would like to have
the
data from the 18,420 participants in the survey/smartphone component
to analyze even if one has to become a MATLAB Wizard to do so (or
get a graduate student who is an apprentice wizard). I'll leave the
fMRI
results and analyses to someone who knows the dark arts used in those
analyses.
I know there are many lessons just waiting to be learned from it and,
we can
all hope, a book or two on the way to explain it to the
non-mathematically
minded among us.
I'm not sure that's such a good thing. Afterall, Fredrickson wrote the
book
"Postivity" that presumably explained all of the necessary math and
stuff
needed to understand why the magic ratio was "real". Too bad she didn't
understand the math.
-Mike Palij
New York University
[email protected]
---
You are currently subscribed to tips as: [email protected].
To unsubscribe click here:
http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=37841
or send a blank email to
leave-37841-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
