Question 1

How do opinion polls work and what does margin of error mean?

Accepted Answer

A poll estimates what a whole population thinks by asking a **sample** of it. The margin of error quantifies how much the sample result might differ from the true figure purely by chance.

**Why a small sample works.** With genuine random sampling, precision depends far more on **sample size** than on population size. This is why a poll of 1,000 people can describe a country of 60 million — and why increasing the sample from 1,000 to 2,000 improves precision only modestly, since the error falls with the **square root** of sample size. Doubling accuracy requires quadrupling the sample.

**Margin of error** is typically quoted at **95% confidence**, and for a sample of around 1,000 it is roughly **±3 percentage points**. That means: if the poll were repeated many times, 95% of the results would fall within 3 points of the true value.

**Two things people consistently get wrong:**

**It applies to each figure separately.** If a poll shows 42% against 38% with a ±3 margin, the **gap** has a larger uncertainty than 3 points — errors on both numbers compound. A 4-point lead in a ±3 poll is not a clear lead.

**It only covers sampling error.** It says nothing about the far larger sources of error described below.

**What the margin of error does not capture:**

**Non-response bias**, now the dominant problem. Response rates have collapsed, and the people who answer may differ systematically from those who do not.

**Weighting decisions.** Raw samples are never representative, so pollsters weight by demographics and past vote. Those choices are judgement calls, and different reasonable choices produce different headline figures — sometimes by more than the margin of error.

**Question wording and order**, which demonstrably move answers.

**Turnout modelling**, which for election polls is often the largest source of error.

**Read the trend across many polls**, not one, and check the fieldwork dates and the pollster.

Question 2

Why do different outlets report the same poll differently?

Accepted Answer

Because a poll produces many numbers, and which one is led on is an editorial choice — before any question of bias arises, the same data genuinely supports several accurate headlines. What a single

Question 3

Why do exit polls sometimes differ from the result?

Accepted Answer

An exit poll surveys voters immediately after voting, and it differs fundamentally from a pre-election poll: it asks about behaviour that has already happened rather than intention. That removes the

Question 4

Why are so many news stories based on surveys?

Accepted Answer

Because they are cheap to produce, easy to write up, and generate coverage reliably — and a substantial proportion are commissioned by organisations with a commercial interest in the result. The

Question 5

What is agenda setting, and how is it different from bias?

Accepted Answer

Agenda setting is the media's influence over what people think about, rather than what they think. It is one of the better-supported theories in media research, and it describes something distinct

Question 6

Does labelling misinformation actually work?

Accepted Answer

Modestly, with conditions — and the research is more encouraging than the pessimistic framing usually suggests, while falling well short of solving the problem. What the evidence supports: Labels

Question 7

Why is science reporting so often wrong?

Accepted Answer

Because a chain of incentives between the laboratory and the reader systematically amplifies certainty and strips caveats — and the errors are introduced at several points, not only in newsrooms.

Related Questions