Skip to content
AP U.S. Government and Politics · American Political Ideologies and Beliefs

Margin of error

Margin of error appears in 6 reviewed AP U.S. Government and Politics questions and sits in American Political Ideologies and Beliefs. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

The margin of error is the range within which the true population value probably falls, given random sampling error. A result of 47 percent with a margin of plus or minus 3 points means the true figure is likely between 44 and 50 percent.

Why it matters

It tells you whether a lead is real or just noise. Smaller samples, such as one age group within a poll, have larger margins of error, so subgroup results are less precise than the overall result.

How the exam tests it

You add and subtract the margin from each candidate's share and check whether the ranges overlap. If they overlap, the race is within the margin of error; if not, the lead is clearly real. You may also be asked what a 95 percent confidence level means.

Don’t confuse it with

Students think a 95 percent confidence level means the pollster is 95 percent sure this one result is right; it actually means that if the same method were repeated many times, about 95 percent of the ranges would contain the true value.

Margin of error in real questions

Example 1Medium
The table shows hypothetical results from a random-sample poll of likely voters in a two-candidate race. Each margin of error is at the 95 percent confidence level. | Group | Sample size | Candidate X | Candidate Y | Margin of error | |---|---|---|---|---| | All likely voters | 1,200 | 47% | 46% | ±2.9 points | | Ages 18-29 | 200 | 55% | 38% | ±7.1 points | | Ages 30-49 | 400 | 47% | 46% | ±5.0 points | | Ages 50-64 | 350 | 44% | 50% | ±5.3 points | | Ages 65 and older | 250 | 46% | 49% | ±6.3 points | Based on the data in the table, which of the following statements is most accurate?
  1. ACandidate X leads among all likely voters by more than the poll's margin of error
  2. BVoters aged 50-64 prefer Candidate Y by a gap too large to be explained by sampling error
  3. COnly among voters aged 18-29 is either candidate's lead clearly beyond sampling errorCorrect
  4. DThe estimate for voters aged 65 and older is more precise than the estimate for voters aged 30-49

Why C is right

A margin of error gives the range within which the true population value probably falls because of sampling error. Among voters aged 18-29, Candidate X's plausible range (55 ± 7.1, or 47.9-62.1%) does not overlap Candidate Y's (38 ± 7.1, or 30.9-45.1%), so the 17-point lead is larger than sampling error can explain. In every other row the two candidates' ranges overlap. Overall, the 1-point gap is well inside ±2.9. For ages 50-64, 44 ± 5.3 reaches 49.3 and 50 ± 5.3 drops to 44.7. For ages 65 and older, the 3-point gap is inside ±6.3. Ages 30-49 differ by only 1 point.
Why the other 3 choices miss
A
Misreads the overall row. X leads by just 1 point (47% to 46%), well within the ±2.9-point margin of error, so the race is a statistical tie.
B
Compares the 6-point gap with the ±5.3 margin as though the margin applied to the gap. Each candidate's share carries its own ±5.3 range, and those ranges overlap (X up to 49.3%, Y down to 44.7%), so sampling error could explain the gap.
D
Reverses how precision works. The 65-and-older group has a smaller sample (250 vs. 400) and so a larger margin of error (±6.3 vs. ±5.0), which makes its estimate less precise, not more.
Example 2Harder
When pollsters report that a survey has a '95% confidence level,' they mean that
  1. A95% of the eligible population was contacted and offered the opportunity to participate in the survey
  2. B95% of respondents gave accurate and honest answers, and 5% are estimated to have provided false responses
  3. Cthe pollster is 95% certain that no sampling bias occurred in the respondent selection process
  4. Dif the same survey were conducted many times using the same sampling method, 95% of the resulting confidence intervals would contain the true population valueCorrect

Why D is right

A 95% confidence level is a frequentist statistical concept: if the survey procedure were repeated many times on independent random samples from the same population, approximately 95% of the calculated confidence intervals would contain the true population parameter. It does not mean the pollster is 95% certain their specific interval is correct — that probability is either 0 or 1 (the true value either is or is not in the interval). This distinction is frequently misunderstood by journalists and the public, who often interpret it as a probability about the specific poll's interval.
Why the other 3 choices miss
A
Incorrect because no poll contacts 95% of the eligible population; polls contact small samples (typically 600-2,000 respondents) and the confidence level refers to the sampling procedure's long-run accuracy, not the fraction of the population surveyed.
B
Incorrect because the 95% confidence level is a statement about the sampling procedure's long-run properties, not about respondent honesty rates; dishonesty or social desirability bias is a separate measurement concern addressed differently.
C
Incorrect because a 95% confidence level says nothing about sampling bias; confidence intervals assume random sampling and are invalid (regardless of stated confidence level) if systematic bias exists in sample selection.

Studied alongside

Other concepts from American Political Ideologies and Beliefs.