If you have ever read a poll that says 52 percent of voters support a plan, plus or minus 3 points, you have already seen a margin of error in action. Knowing how to find margin of error lets you judge whether a survey result is solid or shaky, whether a sample is large enough, and how much weight a headline number really deserves.
This guide explains how to find margin of error for both proportions and means, walks through the margin of error formula one step at a time, and shows worked examples you can copy. You will also learn how confidence level and sample size change the result, how to calculate margin of error in Excel, and which mistakes trip people up most often. Whether you run customer surveys, A/B tests, or an SEO content strategy that depends on audience research, the same statistics apply.
What Is Margin of Error?
The margin of error is a number that describes how far your sample result could be from the true value for the whole population. Because you almost never survey everyone, your sample result is an estimate. The margin of error puts a boundary around that estimate so readers know how precise it is.
It is usually written as a plus or minus value. If a survey finds that 60 percent of customers like a new feature and the margin of error is 4 percentage points, the true share of all customers is likely between 56 and 64 percent. That range is called the confidence interval, and the margin of error is simply half of its width.
Margin of Error vs Confidence Interval vs Confidence Level
These three terms are closely related, and people often mix them up.
- Margin of error: the plus or minus amount added to and subtracted from your sample estimate.
- Confidence interval: the full range created by your estimate plus and minus the margin of error.
- Confidence level: how sure you want to be that the interval captures the true value, most often 90, 95, or 99 percent.
A smaller margin of error means a narrower interval and a more precise estimate. A higher confidence level makes the interval wider, which means a larger margin of error for the same data.
The Margin of Error Formula
Every version of the margin of error formula has the same shape: a critical value multiplied by a measure of how much your estimate varies. The critical value comes from your chosen confidence level. The variation measure is the standard error, which shrinks as your sample size grows.
Margin of error = critical value x standard error
The exact standard error depends on what you are measuring. These are the three formulas you will use most.
Margin of Error Formula for a Proportion
Use this when your result is a percentage, such as the share of respondents who answered yes. Here z is the z-score for your confidence level, p is the sample proportion written as a decimal, and n is the sample size.
MOE = z x sqrt( p x (1 – p) / n )
Margin of Error Formula for a Mean When Sigma Is Known
Use this when your result is an average and you know the population standard deviation, written as sigma.
MOE = z x ( sigma / sqrt(n) )
Margin of Error Formula for a Mean When Sigma Is Unknown
In most real projects you only have the sample standard deviation, written as s. In that case, and especially when the sample is small, use a t critical value with n minus 1 degrees of freedom instead of a z-score.
MOE = t x ( s / sqrt(n) )
How to Calculate Margin of Error Step by Step
Once you know which formula fits your data, the process is the same every time. Follow these steps to calculate margin of error without guesswork.
Step 1: Decide What You Are Estimating
Check whether your result is a proportion (a percentage of people who chose an option) or a mean (an average of a numeric measure such as order value or time on page). This decides which standard error formula applies.
Step 2: Choose Your Confidence Level
Pick the confidence level that suits the decision at hand. A 95 percent confidence level is the default in most surveys, research papers, and polls. Choose 99 percent when a wrong conclusion would be costly, and 90 percent when a rough read is enough.
Step 3: Find the Critical Value
Look up the z-score for your confidence level in the table later in this guide. If you are estimating a mean with an unknown population standard deviation, find the t value for your degrees of freedom instead.
Step 4: Calculate the Standard Error
For a proportion, take the square root of p times 1 minus p, divided by n. For a mean, divide the standard deviation by the square root of n. If you do not know the proportion yet, use 0.5, which gives the largest possible standard error and therefore the most cautious margin of error.
Step 5: Multiply and Convert to a Percentage
Multiply the critical value by the standard error. For proportions, multiply the result by 100 to express it in percentage points. That final number is your margin of error.
Step 6: Build the Confidence Interval
Subtract the margin of error from your estimate to get the lower bound, and add it to get the upper bound. Report both the estimate and the interval so readers see the full picture.
Margin of Error Examples
Worked examples make the formula easier to trust. Each one below follows the steps above.
Example 1: Margin of Error for a Survey Before You See the Results
Suppose you survey 1,000 randomly selected adults and want the margin of error at a 95 percent confidence level. You have not looked at the results yet, so use p = 0.5.
- Critical value: z = 1.96 for 95 percent confidence.
- Standard error: sqrt( 0.5 x 0.5 / 1000 ) = sqrt(0.00025) = 0.0158.
- Margin of error: 1.96 x 0.0158 = 0.031, or about 3.1 percentage points.
This is why national polls often use about 1,000 respondents. That sample size delivers a margin of error near 3 percentage points at 95 percent confidence, which is precise enough for most public opinion questions.
Example 2: Margin of Error for a Sample Proportion
A company surveys 500 customers and finds that 60 percent are satisfied with a new onboarding process. Find the margin of error at 95 percent confidence.
- Sample proportion: p = 0.60, so 1 – p = 0.40.
- Standard error: sqrt( 0.60 x 0.40 / 500 ) = sqrt(0.00048) = 0.0219.
- Margin of error: 1.96 x 0.0219 = 0.0429, or about 4.3 percentage points.
- Confidence interval: 60 percent plus or minus 4.3 gives roughly 55.7 to 64.3 percent.
You can say with 95 percent confidence that between about 56 and 64 percent of all customers are satisfied. The whole interval sits above 50 percent, so a majority is likely satisfied.
Example 3: Margin of Error for a Mean With a Large Sample
A store samples 64 orders and finds an average order value of 80 dollars. The known population standard deviation is 12 dollars. Find the margin of error at 95 percent confidence.
- Standard error: 12 / sqrt(64) = 12 / 8 = 1.5.
- Margin of error: 1.96 x 1.5 = 2.94 dollars.
- Confidence interval: 80 plus or minus 2.94 gives 77.06 to 82.94 dollars.
Example 4: Margin of Error for a Mean With a Small Sample
A team times 15 support calls and finds a mean handling time of 22 minutes with a sample standard deviation of 8 minutes. The sample is small and sigma is unknown, so use a t value. With 14 degrees of freedom at 95 percent confidence, the t value is about 2.145.
- Standard error: 8 / sqrt(15) = 8 / 3.873 = 2.066.
- Margin of error: 2.145 x 2.066 = 4.43 minutes.
- Confidence interval: 22 plus or minus 4.43 gives about 17.6 to 26.4 minutes.
Compare this with the large-sample case. The small sample and the heavier t critical value combine to produce a much wider interval, which is exactly how the math should behave when you have less data.
Common Z-Scores for Each Confidence Level
The z-score is the critical value you use for proportions and for means when sigma is known. These are the values worth memorizing.
| Confidence level | Z-score | Typical use |
|---|---|---|
| 80 percent | 1.282 | Quick internal estimates |
| 90 percent | 1.645 | Early-stage or exploratory research |
| 95 percent | 1.960 | Standard for surveys, polls, and most studies |
| 99 percent | 2.576 | High-stakes decisions and medical or safety research |
Using the data from Example 1, the margin of error would be about 2.6 points at 90 percent, 3.1 points at 95 percent, and 4.1 points at 99 percent. Higher confidence always costs you precision unless you collect more data.
How Sample Size Affects Margin of Error
Sample size is the lever you control most directly. Because n sits under a square root in the formula, the relationship is not one to one. To cut your margin of error in half, you need to roughly quadruple your sample.
| Sample size | Margin of error at 95 percent confidence (p = 0.5) |
|---|---|
| 100 | about 9.8 points |
| 400 | about 4.9 points |
| 1,000 | about 3.1 points |
| 2,500 | about 2.0 points |
This explains why the returns diminish quickly. Going from 100 to 400 respondents halves the margin of error, but going from 1,000 to 2,500 gains barely one point. At some stage, spending more on additional responses gives you very little extra precision.
How to Find the Sample Size You Need
Often you know the margin of error you can tolerate and need to work backward. Rearranging the proportion formula gives the sample size directly.
n = z squared x p x (1 – p) / (margin of error squared)
With p = 0.5 and 95 percent confidence, a target margin of error of 5 percent (0.05) gives n = 1.96 squared x 0.25 / 0.0025, which is about 384.2, so you need 385 respondents. The same approach gives these targets.
| Target margin of error | Sample size needed (95 percent confidence) |
|---|---|
| 5 percent | 385 |
| 3 percent | 1,068 |
| 2 percent | 2,401 |
| 1 percent | 9,604 |
Always round up, never down, and remember that these figures assume a simple random sample and that people actually respond. If you expect a 30 percent response rate, you must invite far more people than the number you need.
Margin of Error for Small Populations
The standard formulas assume your population is very large compared with your sample. When you sample more than about 5 percent of a small population, apply a finite population correction by multiplying the margin of error by sqrt( (N – n) / (N – 1) ), where N is the population size.
For example, if you survey 500 members from a community of 2,000, the uncorrected margin of error at 95 percent confidence and p = 0.5 is about 4.38 points. The correction factor is sqrt(1500 / 1999), or about 0.866, which brings the margin of error down to roughly 3.8 points. Sampling a large share of a small group really does make your estimate more precise.
How to Find Margin of Error in Excel or With a Calculator
You do not need to do the arithmetic by hand every time. Spreadsheets and online tools handle it in seconds, but you should still understand what they are doing so you can catch input errors.
Margin of Error in Excel
Excel includes two functions built for this job. Use CONFIDENCE.NORM when you are working with a z-score and a known standard deviation, and CONFIDENCE.T when the standard deviation comes from a small sample.
- CONFIDENCE.NORM(alpha, standard_dev, size) where alpha is 1 minus the confidence level, so 0.05 for 95 percent.
- CONFIDENCE.T(alpha, standard_dev, size) for the t distribution version.
For Example 3, the formula CONFIDENCE.NORM(0.05, 12, 64) returns about 2.94, matching the hand calculation. For proportions, Excel has no dedicated function, so build it with a cell formula such as =1.96SQRT(p(1-p)/n).
Using a Margin of Error Calculator
A margin of error calculator usually asks for three inputs: confidence level, sample size, and either the sample proportion or the standard deviation. If you leave the proportion blank, most tools default to 50 percent. Use a calculator to check your work, not to replace your understanding of the inputs.
How to Interpret and Report Margin of Error
A number only helps if you explain it correctly. Say that the result is an estimate with a stated confidence level, for example: “60 percent of customers are satisfied, with a margin of error of plus or minus 4.3 percentage points at 95 percent confidence.”
Pay attention to overlapping intervals when you compare groups. If one group scores 54 percent and another scores 51 percent, each with a margin of error near 4 points, the intervals overlap heavily and you cannot claim a real difference. The margin of error for a gap between two groups is larger than the margin for either group alone, so be cautious about headline comparisons.
Also remember that the margin of error covers random sampling variation only. It does not account for a badly worded question, a biased sample, or people who refused to answer.
Common Mistakes When You Calculate Margin of Error
Even careful analysts make a handful of predictable errors. Watch for these.
- Using a z-score with a tiny sample when a t value is the correct choice.
- Forgetting to convert the final answer into percentage points for proportions.
- Using the wrong confidence level, such as 1.645 when the report promises 95 percent.
- Applying the formula to a non-random or self-selected sample, where the number looks precise but means little.
- Treating a margin of error as a guarantee, when a 95 percent interval will miss the true value about 1 time in 20.
- Ignoring subgroups. A survey with a 3 point margin overall can have a margin of 8 points or more for a subgroup of 150 people.
Why Margin of Error Matters for Business Decisions
Margin of error is not just an academic idea. Marketing, product, and operations teams make decisions from sample data every day, and a result without a margin of error can send them in the wrong direction.
A content audit that relies on a survey of readers, for instance, needs enough responses to separate a real preference from noise. The same logic applies when you test two landing pages. If your site attracts traffic that does not turn into inquiries, a small test with a wide margin of error may tell you nothing, and a larger one may reveal a genuine conversion gap.
Feedback programs benefit too. Teams that care about customer relations often send satisfaction surveys, and knowing the margin of error tells them whether a two point drop in satisfaction is a trend or random variation. The same goes for social media monitoring, where sentiment scores drawn from a sample of mentions carry uncertainty that should be reported alongside the number.
If you want collection and follow-up to run without manual effort, you can automate survey collection and follow-ups inside your CRM. And when your data needs to feed a bigger plan, it helps to connect your website, CRM, and SEO so every sample you collect informs one clear strategy.
Frequently Asked Questions About Margin of Error
What is a good margin of error?
For most surveys, a margin of error between 3 and 5 percentage points at 95 percent confidence is considered acceptable. Tighter margins need much larger samples, so the right target depends on how costly a wrong decision would be.
How do I find margin of error without the population standard deviation?
Use the sample standard deviation to calculate the standard error, then multiply by a t critical value with n minus 1 degrees of freedom. For large samples the t value is very close to the z-score, so the two approaches give nearly the same answer.
What is the margin of error for a sample of 1,000?
At a 95 percent confidence level and the most cautious assumption of p = 0.5, the margin of error for a sample of 1,000 is about 3.1 percentage points.
Does population size affect margin of error?
Only when your sample is a large share of a small population. For large populations, the sample size matters far more than the population size, which is why a sample of 1,000 works almost equally well for a city or a country.
Can margin of error be zero?
Only if you survey the entire population, which is a census rather than a sample. Any time you sample, some random sampling variation remains.
Final Thoughts on How to Find Margin of Error
Learning how to find margin of error comes down to a short routine: choose a confidence level, find the critical value, calculate the standard error, multiply, and report the interval. Remember that sample size has the biggest practical influence, that confidence level trades precision for certainty, and that the number only describes random sampling variation.
Apply these steps and your survey results, test findings, and reports become far easier to trust and explain. If you want help turning reliable data into content and rankings, explore our SEO services or book a call with the team.
