Mead 17
issue) and in opinion about issues over time. We can show results in general and for various types of
witnesses (e.g., Administration spokespersons, interest groups, academic experts, etc.).
One might question whether this procedure captures the full national debate on welfare,
because it omits the public. But about welfare, the significant debate has been among elites. As
mentioned above, public opinion about welfare is remarkably unified and stable over time. Most
voters seek to combine generosity toward needy families with demands that welfare adults work. It
is politicians, advocates, and experts who tend to divide over welfare, along any of the four issue
types I have defined. This elite debate is what our coding taps.
One might also question whether the groups called to testify in Congress fully represent elite
opinion. However, witnesses are chosen by partisan staffs working for both sides of the committees.
The speakers are typically prominent figures within the poverty issue network. When I testify in
these hearings, the other witnesses are the same experts I encounter at conferences outside
government. A wide range of opinions is represented. The debate before the committees is much
like the wider national debate, at least at an elite level.
One reason the hearings feature diverse witnesses is that, throughout most of the reform
period, Congress was seriously divided over welfare and other poverty issues. During the 1960s and
1970s, both parties were diverse ideologically. During the 1980s and 1990s, each became
ideologically more cohesive, but there were now greater differences between the parties, and
different parties usually controlled the two chambers. These divisions ensured that a wide range of
witnesses was called to testify. As results show below, most witnesses were left of center in their
politics, even during the conservative 1980s and 1990s. Wide representation appears to have
prevailed even during PRWORA, when a militant new Republican majority controlled the process.
To support this conclusion, Table 1 shows how many of each of the fourteen types of groups we
32
Pamela Winston, Welfare Policymaking in the States: The Devil in Devolution (Washington, DC:
Georgetown University Press, 2002), chap. 3.
Convention