5 May
2007
5 May
'07
8:34 a.m.
this is good but too focused on what you're doing (replicating) and who
you're doing it 'to', rather than on the change in substantive conclusions
you are finding are correct and we all should believe. also, what does the
additional complexity get you? does it reveal additional observable
implications of H&R's theory that they didn't look at and, when you do,
you find it doesn't support or changes the original theory? clarify what
the big picture is in terms of why you're doing what you're doing..
(of course all these things and others you've all written about will
clarify when you write the whole paper, but you can't trust that anyone
will read that far. and even if they do, their views will be organized
and structured by what you write in the abstract)
Gary
On Fri, 4 May 2007, Bilev, Gavril wrote:
Don't Rule Out "Structure" Just Yet
Revisiting Howard & Roessler's "Liberalizing Electoral Outcomes in
Competitive Authoritarian Regimes"
Why do some semi-free elections in competitive authoritarian regimes
produce liberalizing outcomes while others do not? Howard & Roessler
offer a compelling story that emphasizes the ability of the opposition
to form a coalition. However, a reevaluation of their empirical evidence
reveals a more complex relationship between structural preconditions and
the probability of further liberalization. The opposition's behavior is
more constrained by economic growth and prior experiences with
democratization than implied by Howard and Roessler's theory. We
respecify Howard and Roessler's model to disaggregate their predictor
variables' effects. We find that opposition coalitions are less likely
to form: during periods of economic growth, in states with no prior
history of liberal change, and in states with lower levels of opposition
mobilization. Additionally, we examine the counterfactual scenarios
Howard & Roessler pose in constructing their argument. We find their
counterfactuals!
are not reasonable given their data and suggest modifications that better demonstrate
the impact of the above mentioned variables.
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l