WikiSummary, the Social Science Summary Database

 

Page and Shapiro. 1983. Effects of public opinion on policy. American Political Science Review 77: 175-190.

Main Point

Take-home point: Aggregate shifts in public opinion lead to congruent shifts in public policy (at least on visible issues).

Argument: At the aggregate level, public opinion does appear to influence public policy. More often than not, a shift in public opinion is followed by a shift in public policy. These correlations are higher when the public opinion shift is larger, more stable, or more salient. ("Salience" refers to the number of people that answer "I don't know"; fewer "don't knows" means higher salience.) These correlations hold across issue areas.

Caveats

• Pollsters tend to focus on more visible/salient issues. On less important issues, it's entirely possible that the public's voice matters less (cf. Olson 1965).
• In 25-33% of cases, policy moves in the opposite direction as policy: policy responsiveness is far from perfect.
• There is some evidence that policy can move before opinion (reverse causation), although this appears less likely than the reverse.
• This study uses only covariation--it tests only whether policy moves in the same direction as opinion, not whether it moves with the same magnitude.

Empirical Test

• X: They went through survey data spanning decades and looked for instances where a question was asked more than once (possibly with different wording) with a gap of several months or a few years in the middle. In their sample, they included only those questions that actually showed a shift in opinion (X)--thus, there was no control group because they eliminated questions that did not show a significant (>6%) shift in opinion.
• Y: They looked for policy changes during the period starting a couple years before T1 and ending a few years after T2.

Comments and Criticisms

In 1987, the same duo wrote "What Moves Public Opinion," in which they argued that presidents (among others) can change public opinion. Thus, before a policy is implemented, a politician might move public opinion in order to get it passed. In this case, it seems that policy (or at least pre-policy activity) is preceding opinion, which then leads to actualy policy. Elite opinion creates policy, then, not mass opinion; mass opinion would be only an intermediate stage. In the present paper, the authors do not address this possibility. They do consider whether (actual) policy precedes shifts in opinion, but they don't adequately address whether pre-policy activity precedes shifts in opinion.