6
equally likely to side with Republicans as compared with Democrats whereas low status
professionals are more likely to be Democrats.
Cody’s (2003) analysis suffers from two problems: 1) he nonrandomly examined five states;
and 2) he does not look at a legislature that never professionalized. Nonrandom case selection is
a potential problem because it biases estimates of any statistical model based on these cases
(King, Keohane and Verba 1994). It is well known that biased parameter estimates potentially
lead to incorrect conclusions about the population of cases in question (Geddes 1990, King,
Keohane and Verba 1994). As a consequence, Cody (2003) probably received incorrect
parameter estimates as he only looked at five legislatures that professionalized to some degree
between 1960 and 1990. This may have led to his conclusion that is not wholly consistent with
Fiorina’s assumption about the pools of candidates from which the Democratic and Republican
parties draw (1994, 1997 and 1999).
3
Fiorina (1999) builds upon his earlier works (1994, 1997) by looking at whether individual
state ideologies affect the proportion of Democratic legislators within the lower houses of their
state legislatures. To measure the effect of state ideology on the proportion of Democrats, he
uses Berry et al’s (1998) index that measures citizen ideologies within each state from 1960 to
1993. He discovers that more liberal states will have a higher proportion of Democrats within the
lower houses of their state legislatures than conservative states. When he includes state ideology,
his earlier finding that links increased professionalization with more Democrats within state
legislatures’ lower houses still holds (Fiorina 1994, 1999).
Meinke and Haseke (2003) argue that Fiorina (1999) did not examine the effect of term
limits on the proportion of Democrats within U.S. state legislatures. They argue that his omission
is important as term limits partly reverse professionalization and affect candidate incentives in
3
Hereafter, I refer to this assumption as the socioeconomic strata assumption.
Convention