Standard Error of the Mean

According to the Central Limit Theorem, if you take samples from the population to measure any phenomenon, and you take the average of those samples, the histogram of those averages would be shaped like a bell curve. That distribution has the same mean as the population. The standard deviation of that distribution is called the Standard Error of the Mean.

This SEM decreases if you increase the sample size. It will become zero if everyone in the population is measured. It is an indication of how well the sample (of samples) represents the population.

Standard Error of Estimate

A correlation coefficient represents how two variables are related in a linear fashion (if one changes, the other as well). A regression line is the line of best fit. The stronger the relationship between two variables, the less Standard Error of Estimate, and the closer the values to the line.

SEE is also used to determine the Confidence Interval around a predicted score for a criterion.

Standard Error of Measurement

When an instrument is used to to assess a particular aspect in a person, the score usually indicates some level of error. The more reliable assessment will have a lower Standard Error of Measurement. It means that if you repeat the measurement, the scores will be around the person’s true score.