While the GEM Adult Population Survey is, like other surveys and opinion polls, designed to provide estimates of entrepreneurial attitudes, activities and aspirations, such estimates are practically never exactly equal to the ‘true’ rates. Each survey directed at a sample of the target population deals with a “margin of error”, which quantifies uncertainty about (or confidence in) a survey result. GEM usually provides 95 percent confidence intervals to the estimates reported in its global reports. This means that next to the estimates of, for example, Total early-stage Entrepreneurial Activity (TEA) an interval is provided that would contain the true population value at least 95 times for every 100 samples taken from the true population. Larger sample sizes result in smaller margins of error. In addition, in case of rare events (such as the incident of high job-expectation entrepreneurship among the 18-64 population), the relative size of the confidence interval is larger. When analyzing GEM data, 95% confidence intervals can be approximated using the following formula: 

where p is an estimated probability based on the sample and n is the sample size.

They can be visualized in several statistic and graphic software programs. Using Microsoft Excel, the ‘stock’ chart type may be used to visualize estimates along with confidence intervals. The sample size of individuals aged 18-64 years for each economy is indicated in the variable SP1864.