A rant loosely related to [Re: st: comparing policies across countries: multilevel estimation?]

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Thu, 16/12/10, Laura R. wrote:
> doing selection because I assume that is a must with
> employment and wage equations, and worry about multilevel
> later.

That reminds me of one my favourite subjects to rant about:

Each (sub-(sub-(sub-)))discipline has its own issues it 
considers to be crucially important while it happily ignores 
others. Which issues are ingnored and which not is more often 
the result of historical accidents, for example the opinion 
of a particularly influential professor long ago, or (a 
perceived) common error made at some point in the past of the 
discipline. We all have to work within the context of our own 
discipline, so it pays to follow these conventions, but one 
should not take them too seriously. This becomes particularly 
relevant when it comes to comunications across disciplines: 
Too often that just breaks down into a shouting match where 
each side accuses the other of being scientific idiots because 
they ignore problem XYZ (where XYZ differs depending on which 
camp one belongs to). The danger for such a situation is 
particularly increased when "cookbook-style" statistics 
becomes more prominent. In those cases problem XYZ can take
almost magical proportions, and it is hard to have a reasoned 
discusion about magic. On the other hand, a sensible 
discussion about differences in conventions between disciplines 
becomes much easier if one sticks to the basic argument 
underlying most empirical research:

I have a question --> I observed stuff --> I summarized it 
(= statistics) --> I answered my question (for as far as it
is possible)

-- Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/