See what your sample can detect—or calculate the sample required for the difference your study needs to find.
Two different questions: margin of error describes one estimate. Minimum detectable effect describes the comparison between the two entered groups.
Size against the primary outcome and the smallest group you need to report. Weighting efficiency reduces the raw base to the effective base used for inference.
Read the sample-size methodology for formulas, assumptions, unequal groups, and worked examples.
Share the decision, audience, and precision you need. An MX8 Labs researcher can help turn the calculation into a workable study.
Margin of error describes the precision of a single estimate under the calculator's sampling assumptions. At 95% confidence for a 50% outcome, 1,000 respondents give about ±3.1 points; 400 give about ±4.9 points. Halving that margin takes four times the sample.
Comparing two groups takes more sample than estimating one number. The minimum detectable effect is the smallest difference between groups the design is estimated to detect at your chosen confidence and power. If the difference that matters to your decision is smaller than the calculator reports, increase the sample per group before fieldwork starts, not after.
Weighting corrects sample composition but costs precision. The design effect converts your nominal sample into an effective sample size, which determines the calculated margin of error. Enter expected weighting efficiency to estimate the raw sample needed. Actual precision depends on the achieved sample.
The same calculations run inside the MX8 Labs Research Platform, which applies two-stage IPF (raking) weighting and adjusts significance tests for effective sample size. See the sample size and precision methodology and the weighting methodology for the formulas behind each output.