By Jon Zamboni | Updated Aug 30, 2022
ALFSnaiper/iStock/GettyImages
In research, the number of observations—often denoted as n—is pivotal for reliable results. A larger sample gives a clearer picture of the population, reduces the influence of anomalous data, and tightens the margin of error.
Large sample sizes yield more precise averages, highlight true outliers, and shrink confidence intervals, making findings more trustworthy.
Sample size refers to the count of data points collected in a survey or experiment. For example, testing 100 seawater samples for oil residue means a sample size of 100; surveying 20,000 individuals about anxiety yields a sample size of 20,000. While bigger samples demand more resources, the payoff is richer, more robust data.
Calculating the mean— the arithmetic average— becomes more accurate as n grows. Consider measuring height: with 40 participants the mean might be 5 ft 4 in, but with 100 participants it could shift to 5 ft 3 in, offering a more representative estimate. A larger dataset also exposes genuine outliers—values that deviate markedly from the mean—providing valuable insights for further investigation.
Small samples are vulnerable to skewed results. Surveying only four people about political affiliation and finding one Independent voter would incorrectly suggest 25 % of the population is Independent. Expanding the sample dilutes the impact of such anomalies, ensuring statistics reflect reality.
Sample size directly influences a statistic’s margin of error. For binary questions (e.g., car ownership), the margin of error is roughly 100 ÷ √n. A sample of 100 yields a 10 % margin. For continuous metrics (e.g., height), multiply this figure by two times the standard deviation to capture variability. In every case, larger n shrinks the margin, bolstering confidence in the results.
