Sample size

Sample size is the number of units or people included in a sample. It affects the precision of estimates and the ability of a study to detect patterns or effects.

Sampling distribution: The probability distribution of a statistic across repeated samples drawn under the same conditions. Larger samples generally make a statistic’s sampling distribution narrower.
Population (statistics): The full set of units about which a statistical study seeks to draw conclusions. Sample size counts selected units, while conclusions often target this larger population.
Sample size determination: The process of choosing how many units a study should include to meet specified goals and constraints. It turns desired precision or detection ability into a planned sample count.
Sampling bias: Systematic distortion arising when the selection process makes a sample unrepresentative of its target population. More observations do not remove bias introduced by how units entered the sample.
Standard error: The estimated standard deviation of a statistic’s sampling distribution. For many estimates, standard error decreases roughly with the square root of sample size.
Sampling frame: A list or operational representation of units from which a sample can be selected. The frame determines which units can contribute to the reported sample size.
Power analysis: A calculation relating sample size, effect size, variability, significance level, and statistical power. It estimates the sample needed to detect a specified effect with a chosen probability.
Nonresponse bias: Bias that occurs when respondents differ systematically from sampled units that do not respond. A high nominal sample size can conceal information lost through selective nonresponse.
Statistical power: The probability that a statistical test rejects a false null hypothesis under specified conditions. Increasing sample size often raises the chance of detecting a real effect.
Sampling unit: The entity selected at a particular stage of a sampling design. A study’s sample size depends on whether it counts people, households, sites, or another unit.

Linked from 9 pages