Validity vs Accuracy vs Reliability

ABA Measurement Quality: Validity vs Accuracy vs Reliability (and How to Use IOA) This episode explains that data can look clean yet still be wrong, so measurement quality in ABA requires evaluating validity, accuracy, and reliability. Validity means the procedure directly measures a socially significant target behavior, measures a relevant dimension, and samples representative conditions and times; major threats include indirect measurement, measuring the wrong dimension, and measurement artifacts from discontinuous measurement, poor scheduling, or insensitive scales. Accuracy is whether observed values match true values and is prioritized when true values can be checked; reliability is whether repeated measurement of the same event yields the same values and can be high even when inaccurate. When true values are unavailable, reliability is the next best indicator; when permanent products are unavailable, interobserver agreement (IOA) supports believability and helps assess observer competence, drift, and clarity of definitions. The script reviews common IOA methods across event, timing, and interval data and provides a checklist for evaluating measurement systems and threats such as human error, poor design, inadequate training, drift, expectations, bias, and reactivity. 00:00 Clean Data Wrong 00:03 Validity Accuracy Reliability 01:04 Valid Measurement Elements 01:51 Threats to Validity 02:46 Accuracy Basics 03:39 Reliability Explained 04:25 Quality Evaluation Hierarchy 05:26 Human Error Threats 06:31 IOA Purpose Basics 07:25 IOA Calculation Methods 08:45 Final C8 Checklist 09:48 Recap Key Takeaways