Miller, Krosnick, and Fabrigar: The origins of policy issue salience

Miller, Krosnick, and Fabrigar. 2005. The origins of policy issue salience.

Overview

Studies cite the concept of 'salience' frequently, but they have diverse concepts in mind. Miller et al see two concepts: sociotropic salience (e.g. "What is the most important issue facing our country?") and personal salience (e.g. "Which issue is most important to you?"). They argue that these concepts are both conceptually and empirically distinct, and that personal salience plays a much greater role than sociotropic salience in predicting behavior--which is ironic, since politicians appear to respond to sociotropically salient concerns.

Data

The review draws on data from eight separate studies. See the text for descriptions.

Findings: Six Questions, Six Answers

1. Are sociotropic and personal salience empirically distinct?

• Expectation: Yes.
• Finding: Yes.

2. Which has stronger effects on expressed preferences and political activism?

• Expectation: Personal.
• Finding: Yes, you donate more money, write letters to Congress, etc.

3. Which has stronger effects on electoral choices between candidates?

• Expectation: Personal. If something really matters to you, you should vote based on it.
• Finding: Yes. When you interact personal importance with the issue distance btw you and a candidate, the variable is significant. Personal salience matters. National importance does not work this way.

4. Are the stronger effects of personal salience caused by its ability to cause cognitive and emotional engagement on an issue?

• Expectation: Yes. Personal salience promotes accessibility, though, certainty, knowledge accumulation, memory of facts, etc. Like Zaller (1992) said, you're more likely to have things at the top of your head if you think about it more.

5. Is it this increased engagement (H4) that causes the effects in H3 (as an intervening variable)?

• Finding: Some evidence.

6. Does sociotropic salience cause personal salience, making its (sociotropic's) empirical coefficients appear weaker than they really are?

• Expectation: It's possible, but so is the reverse.
• Finding: Some evidence that it goes both ways, but it appears that personal salience affects national importance more than vice versa (but it's significant either way).

Comments and Criticism

Question-Order Problem?

My main concern was with possible question-order effects on studies 1 and 2 creating a false correlation between personal and sociotropic salience.

Sociotropic vs Personal Logic

Prediction # 3 assumes that we think our personal problems are the government's fault (just like all 'pocketbook voting' arguments). But unless we have reason to make this link, why should we vote our personal issues? Perhaps, for this reason, we should expect an interaction between personal and sociotropic salience: Our votes reflect things we care about personally only when we think they are threatened nationally. The authors acknowledge this point in a very unsatisfactory way on p 23.