7
All figures were obtained by an extensive search on Lexis for the published opinions in court
cases occurring in all levels of the state courts of each of the 50 states.
9
We assumed that the
likelihood of a state legislature introducing a proscription of same sex marriage would be
positively related to any current state court decision that simultaneously restricted the celebration
of same sex marriage within that respective state. We coded as one any year with such a decision
presently defining the law in the state. In addition, we assumed, based on the actions of Alaska
and Hawaii, that an active and current pro-same sex marriage court decision might also spur
legislative action on this issue. Similarly, we coded as one any year with such a pro-same sex
marriage decision as the law in the state.
We also introduced into the second regression model two variables to capture the partisan
politics in the state legislature and executive. The data for both variables were obtained from the
figures published in the Book of the States. We assumed that the larger the percentage of the state
house that were Democrats, the less likely it would be that a state would endorse a proscription
on same sex marriage. The percentage of Democrats in the state senate (or its equivalent upper
house) were coded by year for each state, except for the unicameral, non-partisan Nebraska state
legislature where we substituted in the percentage based upon this state’s Congressional
representation. We also assumed a Democratic Governor would be less likely to sign into law
such a proscription and coded a Democratic Governor for the purposes of the regression
equation.
Four additional variables, designed to capture the particular demographics of each state,
were introduced into the model. The data for all four variables were obtained from the decennial
census figures from 1970 through 2000 and used linear interpolation and extrapolation for all
other years. Urban populations have previously been found (Klawitter and Hammer 1999) to be