Measuring Opinion: Polling

You’ll be able to

Explain how scientific polls use random sampling to represent a population
Interpret sampling error, margin of error, and confidence levels
Identify sources of polling error and evaluate poll quality

What makes a poll scientific

A scientific poll estimates the views of a large population by surveying a small sample. The key to accuracy is a random sample, in which every member of the population has an equal chance of being selected — this makes the sample representative. Pollsters may use a stratified design to ensure key subgroups appear in correct proportions. A poll’s quality depends on sampling method, question wording, and sample size. A benchmark poll measures early standing; a tracking poll follows change over time; an exit poll surveys voters as they leave the polls.

Sampling error and confidence

Because a poll surveys only a sample, its result comes with a margin of error (sampling error) — typically expressed as "±3 percentage points." This means the true population value is likely within that range of the reported figure, at a stated confidence level (usually 95%). Larger samples generally shrink the margin of error. When two candidates are separated by less than the combined margin of error, the race is a statistical tie — you cannot confidently say who is ahead. Reading the margin of error correctly is essential to interpreting any poll.

Where polls go wrong

Several problems can bias a poll. A non-representative sample (poor sampling) skews results. Question wording can lead respondents toward an answer. Social desirability bias leads people to give the "acceptable" answer rather than their true view. Non-response bias arises when the people who decline to answer differ systematically from those who respond. Weighting wrong assumptions about who will actually vote (the "likely voter" model) can also mislead. Good polls disclose their methodology so their quality can be judged.

Worked example

A poll of likely voters reports Candidate A at 48% and Candidate B at 45%, with a margin of error of ±3 percentage points. A headline declares "A Leads B." Evaluate whether that headline is justified.

1.Identify the gap: A leads B by 3 percentage points (48% − 45%).
2.Recall what the margin of error means: each candidate’s true support could plausibly be about 3 points higher or lower than reported.
3.Compare the gap to the margin of error: the 3-point lead is within the ±3-point margin, so A’s true support could be as low as 45% and B’s as high as 48%.
4.Conclude: because the lead falls within the margin of error, the result is a statistical tie and the "leads" claim is not statistically justified.

Answer: The headline is not justified. A’s 3-point lead is within the poll’s ±3-point margin of error, meaning the two candidates are in a statistical tie — the data cannot confidently show A ahead of B. Reporting a lead here overstates what the poll can support.

Checkpoint

Why is random sampling essential to a scientific poll?

Watch out

A poll is only as good as its sample. A huge sample drawn unscientifically (like a website’s self-selected online survey) is worse than a small, properly randomized one. Do not assume "more responses" means "more accurate."

Checkpoint

A survey question reads, "Do you support the sensible, common-sense plan to protect families?" This wording most likely introduces which problem?

On the exam

On poll-interpretation items, always compare the candidates’ gap to the combined margin of error. If the lead is smaller than (or within) the margin, call it a statistical tie — that phrase is frequently the intended answer.

Answer the 2 checkpoints as you read.