How should performance data be evaluated for reliability and validity?

• A. Reliability: consistency across trials; Validity: measures reflect the intended construct; use calibration, pilot testing and cross-checks
• B. Reliability only matters
• C. Validity is not important
• D. Use random sampling

Answer: (A)

To ensure performance data are meaningful, focus on both reliability and validity. Reliability means results are consistent across trials or raters; if you repeat the test under the same conditions you should get similar scores. Validity means the measurements actually reflect the intended attribute of performance, not something else. Achieving this often involves calibration of equipment so measurements are accurate, pilot testing to refine procedures and reduce errors, and cross-checks that compare the primary measure with other indicators of the same construct. For example, timing sprint speed should be reliable when devices are calibrated and testers are consistent, and valid when it truly captures speed rather than unrelated factors like reaction time. Relying on reliability alone isn’t enough, and validity alone isn’t sufficient if the data aren’t collected consistently. Random sampling pertains to how data are collected, not directly to ensuring the measurement itself is reliable and valid.