Abstract
Q uality survey data require surveyors to ask the right questions of the right respondents. This requirement is particularly acute for preeminent data sets and templates for other research. The 1984 Ameri_ can National Election Study (ANES) supplemented an item used since 1972 with a new measure of ideological self-location. This paper compares responses to old and new questions and assesses the relative reliability and validity of variables derived from each question. It concludes that data from the traditional question exhibit far more construct validity than those from newer questions, with similar reliability. This conclusion holds for different strata of ideological sophistication. The traditional measure of ideology is better, especially when asked of more sophisticated respondents.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
