What Is a Margin of Error and Why Should You Care?

You've seen it in every political poll, customer satisfaction report, and market research summary: "Results are accurate to within ±3 percentage points, 19 times out of 20."

Most people skim past it. But that small disclaimer carries a lot of weight — and misunderstanding it (or ignoring it) can lead to very wrong conclusions from perfectly valid data.

This article breaks down exactly what margin of error means, where it comes from, how to calculate it, and — most importantly — how to use it correctly when interpreting survey results.

What Is Margin of Error?

Margin of error (MOE) is a statistical range that expresses how much your survey results might differ from the true value in the full population you're studying.

In other words: you didn't survey everyone. You surveyed a sample. Margin of error tells you how far off your sample result might be from what you'd have found if you had surveyed everyone.

Example: Your survey finds that 62% of customers are satisfied with your service. With a margin of error of ±4%, the true satisfaction rate in your full customer base is likely somewhere between 58% and 66%.

That range — 58% to 66% — is what the margin of error gives you. It's an honest acknowledgment that samples are imperfect representations of populations.

Where Does Margin of Error Come From?

Margin of error is rooted in the mathematics of sampling. When you draw a random sample from a population, you're relying on the laws of probability to make that sample representative. But probability isn't certainty — there's always some chance your sample skewed slightly in one direction.

Three factors drive your margin of error:

1. Sample Size

The bigger your sample, the smaller your margin of error. This is the most direct lever you control. A survey of 100 people will have a much larger margin of error than a survey of 1,000 people drawn from the same population.

This is also why sample size planning is so important before you launch — your MOE is largely determined before you collect a single response.

2. Confidence Level

Margin of error is always tied to a confidence level — typically 95% (expressed as "19 times out of 20" in Canadian and some other reporting conventions). This means: if you ran your survey 100 times with different random samples, about 95 of those samples would produce results within your stated margin of error.

Common confidence levels:

Higher confidence = wider margin of error. You can be more certain, but your range gets broader.

3. Population Variability

The more varied your population's opinions are on a topic, the larger your margin of error.

Statisticians account for this using a value called the standard deviation (or, for proportions, assumed variability of 50/50 — the most conservative and most common assumption).

The Margin of Error Formula

For a simple proportion (a percentage result), the margin of error formula is: MOE = Z × √(p(1 − p) / n)

Where:

• Z = the Z-score corresponding to your confidence level (1.96 for 95%)
• p = the proportion you observed (e.g., 0.62 for 62%)
• n = your sample size

Example calculation
You survey 400 customers and find 62% are satisfied. Using a 95% confidence level:

• Z = 1.96
• p = 0.62
• n = 400

MOE = 1.96 × √(0.62 × 0.38 / 400)
MOE = 1.96 × √(0.2356 / 400)
MOE = 1.96 × √0.000589
MOE = 1.96 × 0.02427
MOE = ±4.8%

So your result of 62% satisfied customers has a margin of error of about ±5%. The true satisfaction rate in your full customer base is likely between 57% and 67%.

You can use SimpleSurvey's research tools to plan your sample size upfront and hit your target margin of error before you launch.

How Sample Size Affects Margin of Error

One of the most practical things to understand about MOE is how dramatically sample size affects it — and how quickly the returns diminish.

Here's how MOE changes with sample size at 95% confidence (assuming 50/50 variability, the most conservative estimate):

Notice the pattern: going from 50 to 200 respondents cuts your MOE roughly in half. But going from 1,000 to 5,000 respondents — five times more data collection effort — only reduces MOE from ±3.1% to ±1.4%.

This is called diminishing returns, and it's why most well-designed surveys target a sample size of 400–1,000 respondents. Beyond that, the improvement in precision often isn't worth the additional cost and effort.

Common Mistakes When Interpreting Margin of Error

Mistake 1: Treating MOE as a guarantee

Margin of error is a probability statement, not a guarantee. At 95% confidence, there's still a 5% chance your true result falls outside the stated range. In high-stakes research, that matters.

Mistake 2: Ignoring MOE when comparing two numbers

This is one of the most common errors in reporting survey results.

Example: Candidate A polls at 48%, Candidate B at 44%, with a ±4% margin of error. Many people would call this a 4-point lead for Candidate A.

But when you account for MOE, Candidate A could be as low as 44% and Candidate B as high as 48%. The race could be essentially tied.

This is what pollsters call results being within the margin of error — and it means you cannot confidently conclude one result is higher than the other.

Mistake 3: Applying the overall MOE to subgroup results

If your survey has 1,000 respondents and a ±3% MOE overall, that doesn't apply to subgroups.

If you're analyzing responses from the 80 respondents aged 55+, that subgroup has its own — much larger — margin of error.

This is why surveys that need to report on multiple subgroups require larger overall samples. Each segment needs enough respondents to produce reliable results on its own.

Mistake 4: Confusing margin of error with bias

Margin of error only accounts for random sampling variability. It says nothing about whether your sample was representative in the first place.

If your survey only reached people who opted in through a social media ad, your margin of error might be ±3% — but your results could still be way off because of selection bias.

MOE and bias are separate problems. A small MOE doesn't mean your results are accurate; it means they're precise given your sample. Accuracy depends on how well your sample represents your population.

For tips on reducing bias before it affects your data, see our guide on writing unbiased survey questions.

Margin of Error vs. Confidence Interval

These two terms are closely related and often confused.

1. Margin of error is the ± value: the amount you add and subtract from your result to define the range.
2. Confidence interval is the full range itself: your result minus the MOE to your result plus the MOE.

They're two ways of expressing the same thing. If your result is 62% with a MOE of ±5%, your 95% confidence interval is 57% to 67%.

Most research reports the confidence interval; most news coverage reports the margin of error. Both are valid — just make sure you know which one you're reading.

How to Report Margin of Error Clearly

When sharing survey results, always state:

1. The result — e.g., 62% of respondents said they were satisfied
2. The margin of error — e.g., ±4.8 percentage points
3. The confidence level — e.g., at 95% confidence / 19 times out of 20
4. The sample size — e.g., based on 400 responses

A complete, honest statement looks like this: "62% of surveyed customers reported being satisfied with our service (n=400, MOE ±4.8%, 95% confidence level)."

This gives readers everything they need to evaluate your findings. It's a mark of credible, transparent research.

Ready to Design a Survey With the Right Margin of Error?

Knowing your target margin of error before you launch tells you exactly how many responses you need — which helps you plan your budget, timeline, and outreach.

SimpleSurvey makes it straightforward to build and distribute surveys designed to hit your precision targets.

The Bottom Line

Margin of error isn't fine print — it's one of the most important numbers in your survey results.

It tells you how much to trust a single figure, whether two results are truly different, and whether your subgroup findings are reliable.

Understanding it doesn't require a statistics degree. It requires knowing three things: what confidence level you're working with, how large your sample is, and that MOE only covers random error — not bias.

Get those right, and you'll read survey data — yours and everyone else's — with a much sharper eye.