Statistics in Kinesiology Chapter 13

Reliability Concepts

• Reliability refers to the reproducibility and consistency of measurements.
• Test-Retest Reliability involves administering the same test to a sample at least twice to assess consistency over time.
• Inter-Rater Reliability evaluates the degree to which different raters provide similar scores for the same subject on a given test.
• Intra-Rater Reliability assesses the ability of the same rater to give consistent scores across multiple administrations of a test.

Measurement Error & Components

• Reliability quantifies measurement error, indicating the accuracy and precision of test scores.
• Reliability Theory emphasizes that all measurements contain errors, with observed scores comprising a true score and an error component.
• The True Score Component represents the average score from infinite trials, while the Error Component is the difference between the observed score and the true score.

Types of Error

• Random Error is unpredictable and typically averages to zero over time, with variations in scores that cancel each other out.
• Measurement tools to quantify reliability include repeated measures ANOVA to derive variance needed for reliability coefficients.

Intraclass Correlation Coefficient (ICC)

• The ICC formula is calculated by taking the ratio of true score variance to total variance, yielding values from 0.0 (all error) to 1.0 (all true variance).
• The ICC can be estimated using between-subject variability and error terms from repeated measures ANOVA, serving as an index of true score variance.
• Models 1, 2, and 3 in ICC calculations vary based on how error components are treated, with Model 1 lumping them together and Models 2 and 3 separating them.

Evaluating ICC & Systematic Errors

• When assessing reliability, an increase in between-subject variability improves ICC values.
• Systematic errors can emerge, indicated by significant mean differences across trials, requiring further scrutiny during testing.
• Practitioners should account for systematic errors in interpretation, such as through practice sessions to mitigate these effects.

Standard Error of Measurement (SEM)

• SEM provides an absolute measure of reliability and indicates test precision.
• Two common SEM formulas exist, with the most prevalent reflecting the relationship with ICC.
• SEM estimation is insensitive to between-subject variability, focusing on measuring precision across trials.
• Practitioners can use SEM to construct confidence intervals around individual test scores, assessing boundaries of true scores.

Application in Practice

• SEM aids clinicians and coaches in determining the minimal detectable change post-intervention, allowing for assessment of performance improvements.
• Practitioners should consider a three-layered approach: utilizing repeated measures ANOVA for systematic error assessment, calculating ICC, and determining SEM to quantify measurement error comprehensively.

Test your knowledge of key terms in Chapter 13 of Statistics in Kinesiology. This flashcard set covers important concepts such as reliability, test-retest reliability, inter-rater reliability, and intra-rater reliability. Perfect for students looking to reinforce their understanding of statistical principles in kinesiology.