A survey’s sample size is the number of respondents you need for statistically reliable results. For most surveys, that means about 385, at a 95% confidence level with a 5% margin of error. Sample size matters double for a data journalism or PR campaign, because respondent count also tracks with the coverage a study earns; across 800 Fractl campaigns, surveys of 1,000-plus respondents earned a median of 13 brand mentions, more than double the 5 that surveys under 500 earned.

Our survey sample size calculator sizes your survey and shows how that respondent count has tracked with coverage:

How To Calculate Survey Sample Size

1. Set your confidence level. This is how sure you want to be that the results reflect your target population. A 95% level of confidence is standard and corresponds to a z-score of 1.96 (z-score refers to how many standard deviations a specific response or group average is away from the overall mean of all responses).
2. Choose a margin of error. Also called the confidence interval, this is how much sampling error you’ll accept, usually 5 percentage points either way.
3. Estimate your population proportion. When you don’t know how responses will split, use 50%, the most conservative expected proportion, which produces the largest minimum sample size.
4. Add your population size if it’s small. For a finite population, a finite population correction lowers the number of participants you need; for a large or unknown population, skip it.

Together, those inputs determine sample size. A typical survey needs about 385 respondents, and tightening the margin of error to 3% pushes that past 1,000. The formula assumes simple random sampling, so the further your recruiting strays from a random sample, the more caution the result deserves. If you’re measuring a number like income or age (continuous data) rather than a yes-or-no split, you enter an estimated population standard deviation instead of a proportion.

What Sample Size Earns Press Coverage

A sample size that satisfies a statistician is only half the question for a PR campaign. The other half is whether the respondent count helps the study get covered. We read the respondent count from the published methodology of 800 Fractl survey campaigns and joined each one to the brand mentions it earned. Median brand mentions climb with sample size.

There are two caveats. First, topic and promotion move coverage far more than sample size does, so read the trend as directional, not a promise. Second, almost every campaign in the 1,000-plus group ran between 1,000 and 2,000 respondents, so the trend flattens above that; a 5,000-person survey doesn’t earn proportionally more coverage.

We report medians rather than averages because a few viral studies pull an average up and overstate what a typical survey earns.

Sizing a Survey That Gets Coverage

For market research meant to inform an internal decision, the statistical minimum is enough: hit your confidence level and margin of error and stop. For a study meant to earn media, a sample of 1,000 or more respondents is the safer minimum, because it clears the reliability bar journalists and fact-checkers look for and matches the 1,000-plus respondent range that earned the most coverage in our data.

It also lets you segment and analyze survey results by generation, region, or income tier, which is where the quotable angles come from. This is different from an academic cohort study or case-control study, where the required sample size depends on the effect you’re trying to detect and the statistical power you need, not on whether the finding makes news.

Turn a Survey Into Coverage That Compounds

The right sample size clears the credibility bar, but the story still has to be worth covering, and building surveys that earn national pickup is what we do at Fractl. Want to produce data that publishers want to share with your shared audience? Partner with Fractl’s digital PR team.

Survey Sample Size Questions