ABSTRACT
The GOTV and participation experiments reported in the last chapter have led to a greater understanding of the political behavior of citizens, especially in response to external influences. These field experiments have a made a significant contribution to an important branch of behavioral political science. Yet, much of political science is concerned with the operation of institutions and the representation of interests within the political system, which would appear to exclude experiments as a way to offer answers to these research questions. This limitation is partly – as discussed earlier in Chapter 2 – because political institutions are usually hard to manipulate and are often in insufficient numbers to sustain statistical modeling. However, actors within institutions, the elected representatives, do exercise an important role in structuring those institutions to make policy choices. They receive and act on information from their environment so they can be manipulated. Their numbers are not as great as the citizens who are the targets of GOTV campaigns, but there are many elected representatives and even more bureaucrats to study. By examining legislator responses to constituent demands, it is possible to contribute to debates about equality and representation (see Butler 2014) from how representatives respond to interest groups and campaigns or are held accountable when carrying out institutionally defined roles. It is possible to get insights into decisionmaking and on how leaders balance interests. With bureaucrats, it is possible to test for their responsiveness. Even if the institution itself is not varied, experiments can help political scientists understand how the institution works in a different way from much observational research. As Grose argues, there is “vast potential for field experiments in the study of political institutions” (2014, 356). One question that emerges is whether issues of sample size and manipul-
ability limit the range of these experiments. The bigger question is whether the switch from relatively powerless citizens – who might write a complaining email or make a telephone call – to more powerful politicians and bureaucrats has implications for the range and type of experiments that can be done. It is also not possible to vary important signals that politicians and bureaucrats respond to, such as external crises, media scandals, strategies
from opposing political parties, Supreme Court decisions, competition with other jurisdictions, and the like. On the other hand, experimenters have been ingenious in discovering new way to test for elite responses, which the following sections of text will show.
