Captain Amazing: Let me expand on the coincidence —> correlation —> causality chain that I think Lott has a problem with. For now, I’ll stick to general observations about Lott’s work (possibly this weekend I’ll have time to get into specifics regarding the Women’s Suffrage paper).
(NOTE: IANA statistician, either by profession or hobby. My observations are driven by skepticism, resistance to partisan bullshit and a neurotic distrust of easy answers.)
Coincidence: The starting place for social observations. What broad social actions were taken during a specific time period, and what specific changes occurred in social phenomena concurrently and subsequently to those actions? Lott generally does a thorough job of examining the application and scope of the specific social or legislative actions which are the focus of his studies. He is less thorough in his examinations of changes in social phenomena, preferring to focus on the specific effects he has presupposed to be present.
Correlation: Do the coinciding social variables relate to each other in a way not expected on the basis of chance alone? Lott’s methods of establishing correlation between actions and specific phenomena are fairly solid. Where he fails (and this is what throws up the giant “red flag” for me) is in his inattention to other relevant correlations. Lott gives a wealth of data establishing the relationships he wants to present, but is… less dogged (cursory or disinterested might be a less diplomatic way to put it) in his pursuit of relationships not included in his thesis. Further, Lott seems to examine other social actors only to the degree necessary to show their inclusion; frequently he will underestimate or “miscalculate” the strength of those correlations.
Causality: Is there a causative link between the action and the phenomena? Thanks to John Stuart Mill, we can establish causality through four general methods.[ul][li]Concomitant variation - if a phenomenon varies in any way whenever another phenomenon varies in a particular way, a causal relationship exists. (Lott loves this method.)[]Difference - if a phenomenon occurs in one instance, but not in another, and the two instances differ in only one circumstance, that circumstance is either the effect, cause or concurrent effect of the cause of the phenomenon.[]Residues - rule out those phenomena known to be the effects of other antecedents and the phenomena which remain are the effects of the remaining antecedents.Agreement - if two or more instances of a phenomenon have only one circumstance in common, that circumstance is either the cause or an effect of the phenomenon.[/ul]Note that these methods establish only that a causal relationship exists. When one method is used in isolation from the other methods, it is impossible to determine direct causality. Lott, by relying so heavily on concurrency and concomitant variation to establish causal relationships, can never show whether the variables are common effects, common causes or interrelated in a more complex manner; he can merely establish that a relationship exists.[/li]
Remember, sociological research is meat and potatoes for legislators. Once a “scientific study” establishes a causal link between a closely held tenet of a political philosophy and a widely desired social effect, moral justification can be given for any number of proposals. It doesn’t matter how flawed a published paper actually is; if Congressman Joe can say “studies show…” he’s well on his way toward garnering support for whatever ill-considered bill he happens to be sponsoring.
Whew! I’ll try and be less wordy in the future, but I’ll be back in a day or so to deal with Lott re: suffrage.
[/quote]
Ex Tank: What gives? You gave a link to an article which RTFirefly had spent some effort a few months ago debunking. He felt it prudent to mention that fact, and to assert (politely and briefly) that he doesn’t consider Crook’s article very creditable. I don’t understand why this was a “dig” at you any more than Tominator2 referencing Lambert in response to Lott was a “dig” at Freedom.
[/quote]
Freedom: I see no reason to believe “shall issue” laws have any negative effect on public safety (as long as strict criteria are followed); I just think it’s irresponsible to suggest that these laws “prevent” crime.
