One of these criteria is the survey’s margin of error. Since surveys only question a sample of a larger population that is being studied – whether that population is a single city, an entire country or something else – the margin of error describes the estimated range within which we would expect the exact answer to fall. (The results we would have gotten if we had surveyed everyone in that larger population is the “true population value.”) For example, if a survey has a margin of error of plus or minus 3 percentage points at the 95% confidence level, that means we can expect the result to be within 3 percentage points of the true population value 95 out of 100 times.
http://www.pewresearch.org/fact-tank/2018/08/03/when-writing-about-survey-data-51-might-not-mean-a-majority/