Measuring constituency ideology in US house districts: A top-down simulation approach

One of the most intractable problems associated with studying representation in the U.S. House of Representatives involves the measurement of district-level constituency opinion. In measuring constituency opinion in House districts, scholars have relied on a number of alternative approaches, including the use of demographic variables, small-sample estimates of public opinion, presidential election results, referenda data, and "bottom-up" simulated opinion. In this article we develop an innovative "top-down" simulation of House district opinion that provides more reliable and valid measures of House district ideology. We model state-level ideology (as measured by Erikson, Wright, and McIver 1993) as a function of various demographic and political variables found at both the state and House district levels, and then use the estimates from the state-level model to generate predicted ideology scores for each House district during the 1980s and 1990s. Our findings suggest that the top-down simulated measure is a valid indicator of House district ideology that can be used in a number of research venues.