Erikson, Wright, and McIver: Statehouse democracy

Erikson, Wright, and McIver. 1993. Statehouse democracy: Public opinion and policy in the American states. Cambridge: Cambridge University Press.

In Brief

X: Mean ideology in a state

Y: Mean policy output

Thesis: State politics are responsive to state public opinion. Generally, state political outcomes correlate very strongly with state ideology.

Ch 2: Measurement

Data is culled from 13 years' worth of regular telephone polls by the New York Times and CBS. The data are pooled into a single sample. Lots of tests to show that the data are reliable and valid.

Ch 4: Public opinion and public policy

Public opinion (X) correlates very strongly with policy outcomes (Y), therefore the authors conclude that state politics are responsive. The effect of X "overwhelms" the explanatory power of the controls. (And two stage least squares shows that it is not policy that drives ideology.)

• Critique: They haven't proven anything. If a state's population is more conservative, then you would expect that most candidates/politicians would also be more conservative (on average); thus, even if they completely ignore the electorate, you would expect them to produce more conservative political outcomes. Of course, it's incredibly difficult to measure representativeness. It's hard to do better than this study, which simply shows that policy outcomes reflect the state population's preferences, which is an interesting thing to prove.

Ch 5: State parties and state opinion

MAIN POINT: Party elite ideologies are a reflection of state ideologies. Party activists tend to be more extreme, but electoral elite need to moderate in order to win votes. State public opinion (X) correlates well with elite ideology. Both Democrat and Republican elites are more liberal in more liberal states, although Democrat elites are consistently left of the population and Republican elites are consistently right of it (Duh?).

COMPETING PROCESSES: Responsible Parties vs. Downs. RP demands ideologically driven parties; Downs demands electorally driven politicians. As it turns out, both models are correct: activists follow RP, elected officials follow Downs. Each acts to pull the party to its particular location on the "liberalism continuum."

CONCLUSION: Earlier tests have focused too much on where class conflict appears, and not enough on the policy concerns (i.e. state ideology) of the various state constituencies. As it happens, the location of political conflict in the state government (as evidenced by the liberalism midpoint and the extent of polarization) is determined by the characteristics of public opinion and ideology.

CRITICISM: This simply reinforces my concern from chapter 4.

Ch 7: Test's of Elazar's subcultures

There are two hypotheses in the lit about representative government: either representatives have a vision that they seek to implement ("responsible party government"), or else they change their platform to converge on the median voter (Downs). (The discuss earlier in the book that both are true, to a degree). If Elazar is right, we would expect Moralistic states to have "responsible party government" and Individualist states to look more like Downs said. As it turns out, Elazar was right. Moralistic states have stronger differences between the parties (and elected officials) than Individualist and Traditional states. And elites in Individualist states are more responsive to public opinion than elites in Moralist or Traditionalist states.

Ch 8: Partisanship, ideology, and state elections

CONTRIBUTION: The chapter highlights both the importance and the limitations of state partisanship as a predictor of state election outcomes at different levels for president, governor, and senator.

RESEARCH QUESTION: To what extent are statewide election outcomes determined by state levels of Democratic versus Republican partisanship?

HYPOTHESIS: The impact of state partisanship is different in different types of elections. State partisanship is a strong predictor of state election outcomes; however, in presidential elections, state electorates appear to discount that portion of their partisanship that is traceable to the ideological locations of their state parties. Furthermore, state-level voting in recent presidential election is more a function of state ideological identification than of state party identification.