Reliability refers to the reproducibility of test results in repeated trials. It is crucial for precision and tracking changes in measurements. Key measures include within-subject variation, systematic mean change, and retest correlation. Within-subject variation is best represented by the standard error of measurement, often expressed as a coefficient of variation. Limits of agreement, the 95% range of individual measurement changes, are also used but have limitations. Systematic mean changes, like learning or fatigue, must be accounted for. Retest correlation is complex due to sample heterogeneity. Reliability is essential for decision-making, test comparison, sample size estimation, and individual difference assessment. A minimum of 50 participants and 3 trials are needed for reasonable precision. Complex designs are often misused. The typical error is a better measure than limits of agreement for most applications. Within-subject variation is the most important reliability measure, affecting precision. It is calculated as the standard deviation of repeated measurements. Typical error is the standard deviation of an individual's repeated measurements. It is affected by biological and technical factors. Limits of agreement are 95% ranges of individual measurement changes but are biased and depend on sample size. Change in the mean is a reliability measure reflecting systematic changes. It is influenced by learning, motivation, or fatigue. Retest correlation measures the agreement between trials but is sensitive to sample heterogeneity. The typical error is a better measure than retest correlation for most applications. Reliability is used for monitoring individuals, comparing tests, estimating sample size, and assessing individual differences. The typical error is a better measure than retest correlation for most applications. Sample size estimation depends on the typical error and retest correlation. The typical error is the standard deviation of repeated measurements, while retest correlation is the correlation between trials. Individual differences in response to treatments can be estimated using reliability analysis. The typical error is the standard deviation of repeated measurements, while individual differences are the standard deviation of true effects. Analysis of variance and mixed models can be used to estimate individual differences. Reliability studies should have sufficient participants and trials to ensure precision. The typical error is estimated from the standard deviation of difference scores. Analysis of variance and linear models are used to estimate typical error. Heteroscedasticity, where typical error varies between participants, must be addressed through transformations. Reliability studies are essential for assessing test and equipment performance. The typical error is a better measure than retest correlation for most applications. Reliability studies should be designed with sufficient participants and trials to ensure precision. Analysis of variance and linear models are used to estimate typical error. Heteroscedasticity must be addressed through transformations.Reliability refers to the reproducibility of test results in repeated trials. It is crucial for precision and tracking changes in measurements. Key measures include within-subject variation, systematic mean change, and retest correlation. Within-subject variation is best represented by the standard error of measurement, often expressed as a coefficient of variation. Limits of agreement, the 95% range of individual measurement changes, are also used but have limitations. Systematic mean changes, like learning or fatigue, must be accounted for. Retest correlation is complex due to sample heterogeneity. Reliability is essential for decision-making, test comparison, sample size estimation, and individual difference assessment. A minimum of 50 participants and 3 trials are needed for reasonable precision. Complex designs are often misused. The typical error is a better measure than limits of agreement for most applications. Within-subject variation is the most important reliability measure, affecting precision. It is calculated as the standard deviation of repeated measurements. Typical error is the standard deviation of an individual's repeated measurements. It is affected by biological and technical factors. Limits of agreement are 95% ranges of individual measurement changes but are biased and depend on sample size. Change in the mean is a reliability measure reflecting systematic changes. It is influenced by learning, motivation, or fatigue. Retest correlation measures the agreement between trials but is sensitive to sample heterogeneity. The typical error is a better measure than retest correlation for most applications. Reliability is used for monitoring individuals, comparing tests, estimating sample size, and assessing individual differences. The typical error is a better measure than retest correlation for most applications. Sample size estimation depends on the typical error and retest correlation. The typical error is the standard deviation of repeated measurements, while retest correlation is the correlation between trials. Individual differences in response to treatments can be estimated using reliability analysis. The typical error is the standard deviation of repeated measurements, while individual differences are the standard deviation of true effects. Analysis of variance and mixed models can be used to estimate individual differences. Reliability studies should have sufficient participants and trials to ensure precision. The typical error is estimated from the standard deviation of difference scores. Analysis of variance and linear models are used to estimate typical error. Heteroscedasticity, where typical error varies between participants, must be addressed through transformations. Reliability studies are essential for assessing test and equipment performance. The typical error is a better measure than retest correlation for most applications. Reliability studies should be designed with sufficient participants and trials to ensure precision. Analysis of variance and linear models are used to estimate typical error. Heteroscedasticity must be addressed through transformations.