Public Opinion and Political Beliefs

This is Unit Four of American Government: public opinion and political beliefs. The capstone describes the unit as a loop, and it is a useful shape to carry. Opinion gets formed by upbringing and experience. It gets measured by polling. And then representatives argue about what they owe it. The middle of that loop is where this unit does its most transferable work. A third of the entries are about how polls are built and how they fail — and those techniques apply to any survey anyone ever quotes at you, political or not. The single most useful item is an error entry, and it is worth stating now. A larger sample does not fix a biased one. Bias survives at any size — increasing the count only makes the wrong answer more precise. The unit ends on a question it declines to answer: whether an official should follow the measured public or lead it. That is not something evidence settles. By the end you will be able to do five things. One: name the agents of socialisation, and say why a person's politics is a residue. Two: tell an ideology from a party. Three: say why a random sample beats a larger one. Four: read a poll by four numbers and one sentence, before its headline. Five: state the two models of representation, and why the dispute is not empirical.

Three entries on formation. The term itself, the process, and the sources doing the teaching — and the third contains the useful idea, which is that the sources disagree with each other.

The term.

Gov 161 Public Opinion

Popular views about a person, an event, or an idea. It is a collection — there is no single national mind holding a position.

The process.

Gov 162 Political Socialization

How people are trained to understand and take part in political life, starting very young. Note what it is not — training rather than destiny, with adult experience continually revising what was taught.

The sources.

Gov 163 Agents of Socialization

Family, schools, friends, religious organisations, and media. And they conflict with one another — a person's politics is what is left after competing teachers have all had their turn.

Four entries separating two things that travel together and are not the same. A belief structure about what government should do, and an organisation that runs candidates — they correlate strongly, which is exactly what makes the confusion easy.

The belief structure.

Gov 164 Political Ideology

An organised set of beliefs about what role government should play in society. A structure of belief — not a membership in anything.

How they are arranged.

Gov 165 The Ideological Spectrum

Ideologies ordered by how much government involvement in the economy and society they favour. The two familiar American labels are positions on that spectrum rather than the ends of it.

And the conflation.

Gov 166 ERROR: "Ideology and Party Are the Same"

A belief system is not a party. They usually line up, and that correlation is what makes treating them as one thing tempting and wrong.

The attachment.

Gov 167 Party Identification

A standing attachment to a party, formed early and frequently passed down through the process from the last topic. It is an attachment rather than a registration record — the two can differ.

Three entries on polling, and the third is the most useful item in the unit. It corrects an intuition nearly everyone has about what makes a poll trustworthy — and the intuition is not merely imprecise, it is backwards.

What a poll does.

Gov 168 Scientific Polling

Careful questions put to a representative random sample, to estimate what a whole population thinks. The sample estimates the population and never is the population — everything else follows from that.

What makes it work.

Gov 169 Random Sampling

Every member of the population having an equal chance of being chosen, which is what makes the responses resemble the whole. Representativeness comes from how people were selected, not from how many.

And the error.

Gov 170 ERROR: "A Bigger Sample Beats a Random Sample"

A larger sample does not beat a random one — a biased sample stays biased however large it gets. Size narrows the margin of error only after randomness has done its work.

Three things that move a poll's number before anybody reads it. A range around the figure, the wording that produced it, and a way respondents can mislead the measurement without lying about anything they intend to do.

The range.

Gov 171 The Margin of Error

The band around the reported figure inside which the true value probably sits. Small margins mark quality — a margin of eight points is far too wide to be useful.

The wording.

Gov 172 Question Wording

Questions have to be clear and unloaded, because how something is asked moves the answers. Reading the question is part of reading the poll.

And the respondents.

Gov 173 The Bradley Effect

A gap between what a poll found and what an election produced, caused by people giving the socially acceptable answer rather than their intention. Polls record what was said — they cannot reach past that.

Two models drawing opposite conclusions from the same fact of having been elected, and then a pattern in approval numbers — stated carefully as a tendency rather than a rule, which is the right way to state it.

The first model.

Gov 174 The Delegate Model

A representative in office to voice the people, voting the way they would vote. The election is a set of instructions on this reading.

The second.

Gov 175 The Trustee Model

A representative elected to exercise independent judgment for the people. The same election, read as a grant of discretion — and the disagreement between the two is a standing one the course does not resolve.

And the pattern.

Gov 176 The Presidential Approval Pattern

Presidential approval typically starting high and falling across a term, moving with events at home and abroad. It is a tendency observed across terms rather than a rule any particular term has to follow.

One checklist, and the order it imposes is the point. Four numbers and one sentence get checked first, and a poll can be rejected on those before its headline figure is ever considered.

The checklist.

Gov 180 Reading a Poll

Check the margin of error, the dates, how many people responded, and which population was sampled — then read the question itself. A poll failing those is set aside before anyone gets to the number everybody quotes.

The unit closes on the loop, and then names a question at the end of it that no amount of measurement answers — which is the correct place for a course to stop.

The capstone. A loop, and a question.

Gov 181 The Measured Public

Socialisation forms opinion, polling measures it, and representation argues over what is owed to it. Whether officials should follow the measured public or lead it is a question about what ought to be — the course maps it rather than answering.

Five things you can now do. You can name the agents of socialisation — family, schools, friends, religious organisations, media — and say that they conflict, so a person's politics is what survives competing teachers. You can tell an ideology from a party: one is a structure of beliefs, the other an organisation, and their strong correlation is what makes the conflation easy. You can say why a random sample beats a larger one — bias survives any size, and size narrows the margin only after randomness has done its work. You can read a poll by its margin, its dates, its respondent count, its population, and its question, before you look at the headline number. And you can state the two models of representation, and say why choosing between them is a question about what ought to be rather than one evidence settles.