Majoritarian model v. Pluralism.

Is public opinion stable, clear and informed?

enough for majoritarian model to work?

If only elites are informed, pluralism.

US has weak parties.

V.O. Key: echo chamber campaign

voter knowledge depends on candidates.

Free & Cantril: 2 consensi

ideological v. programs.

Are opinion polls accurate?

Can a few respondents predict a large number of votes?

Winston Churchill, 1946.

Dewey 'defeats' Truman, 1948.

Gallup final (Friday) presidential polls.

Push Poll v. objective survey.

Leading questions versus neutral questions.

Focus Group v. national sample.

Testing phrasing versus broad themes.

Snapshot versus Time Series (panel study).

Benchmark versus Tracking polls.

On-site versus telephone polls.

Likely voter /turnout problem.

Exit Poll.

Validity.

Reliability.

Structured, random sample.

N = sample size.

Margin of sampling error

at 95% confidence level.

N=400, +/- 6% points

N=600, +/- 5 % points

N=1,500, +/- 3% points

Timing: "if election were tomorrow..."

Doorstep opinion problem.

Dichotomous questions.

One tailed versus two-tailed questions.

Response categories.

Ecological problem.

Comparative polls:

translation, cultural problems.

Even 'liberal' changes meaning.

Legitimacy of polls before election.

Americans poll well:

honest & open, fairly high trust.

Unaware of ideological schisms.

Can't identify by social class.

British polls 1992

strongly predicted wrong party.

Voters lied.