Meaning and measurement in cross-national research on satisfaction with democracy

The stability of a democratic nation has long been thought to rest on its level of legitimacy among the mass public. Yet, measurement of such support has been characterized by considerable confusion. One key element in that confusion is the heavy reliance over the past 20 years on data from a survey item that measures respondents' levels of satisfaction with democracy. Data from this item have been analyzed in numerous studies of political support. This research has proceeded despite the existence of substantial disagreement regarding what dimension or dimensions of support the item measures. In an effort to resolve this ambiguity, we examine the conceptual and empirical properties of the item in question. The analysis draws on original surveys conducted in 1999 in Romania and El Salvador and on data from the 1997 Latinbarometer. Results reveal that the satisfaction with democracy item taps multiple dimensions of political support and that the substantive content represented by the item varies across both individuals and nations. We argue that these empirical characteristics limit the capacity of analysts to derive meaningful inferences from study of this item and that, until clarification of the measurement issue is obtained, progress in identifying predictors of democratic stability will be slowed.