RBT · Data Collection and Graphing · 5 min read
Data Reliability and Measurement Errors
- Distinguish between measurement accuracy, measurement reliability, and measurement validity, including why reliable data are not automatically accurate.
- Identify common causes and clinical consequences of missed observations, duplicated observations, and observer drift.
- Describe the RBT's ethical role during data collection, interobserver agreement checks, and handling measurement errors under supervisor guidance.
Foundations of Quality Measurement: Accuracy, Reliability, and Validity
In applied behavior analysis, high-quality data form the bedrock of every clinical decision. As a Registered Behavior Technician (RBT), your primary role during data collection is to record target behaviors objectively as defined in the behavior plan. To understand why exact recording is so essential, you must understand three core measurement concepts: accuracy, reliability, and validity.
Measurement accuracy occurs when the recorded numerical value matches the true state of the behavior. For instance, if a client engages in vocal protest exactly four times during a session, an accurate record reflects exactly four instances. Measurement reliability refers to the consistency of measurement across repeated observations. If two observers watch the same event, or if one observer measures the same response repeatedly under identical conditions, reliable measurement produces the same numerical results.
Measurement validity means the procedure measures the target behavior and the dimension relevant to the goal, for example, duration when the goal concerns how long a behavior lasts. However, it is critical to realize that reliable data are not automatically accurate. An observer can consistently record an incorrect measurement over and over again. For example, if an RBT consistently marks an incorrect behavior code every time a student speaks out, the resulting data are highly consistent (reliable) but completely wrong (inaccurate).
| Measurement Property | Definition | Key Takeaway for Practice |
|---|---|---|
| Accuracy | The recorded numerical value matches the true state of the behavior. | Captures the true occurrence of what actually took place. |
| Reliability | The consistency of measurement across repeated observations. | High consistency does not guarantee that the recorded data are correct. |
| Validity | The procedure measures the target behavior and the dimension relevant to the goal. | Ensures the measurement procedure directly targets the relevant behavior. |
Common mistakes
- Guessing or estimating missed data after a distraction or device failure; RBTs should document the observation gap as directed by the written procedure and tell their supervisor rather than inventing numbers.
- Reconstructing data from memory at the end of the day instead of recording immediately following the occurrence of the behavior.
- Assuming that high interobserver agreement proves data are accurate, even though two observers can consistently apply the same incorrect standard.
- Altering intervention steps or changing behavior plans independently when data appear incorrect, which exceeds the RBT scope of practice.
Key points
- Accuracy means recorded data match what occurred; reliability means measurement is consistent; validity means the procedure measures the target behavior and a dimension relevant to the goal.
- Reliable data are not automatically accurate because an observer can consistently record an incorrect measurement.
- Missed observations falsely lower reported frequency, while duplicated observations artificially inflate frequency or rate.
- Never invent missed observations; document gaps according to the written data-collection instructions, and record observed behavior as promptly as possible.
- Observer drift occurs when an observer gradually and unintentionally changes how they apply a definition; reviewing written definitions and attending supervisor check-ins can help reduce it.
- Interobserver agreement requires independent recording without communicating; high agreement shows consistency but does not guarantee accuracy.
- BCBAs rely on RBT data to make clinical decisions; inventing data violates the RBT Ethics Code, and admitting data errors protects client welfare.
Practice this section
Explained RBT Data Collection and Graphing questions, no account needed. RBT Data Collection and Graphing practice test
The full lesson continues with 3 more sections with exam tips, audio in English and Spanish, 5 Check yourself questions and 20 flashcards.
- 1 month
$44.99$22.49 - 3 months
$131.99$65.99 - Lifetime
$269.99$134.99
One payment, no subscription.
See RBT plansAlready a member? Log in
Sample lesson: Operational Definitions of Behavior25 practice questions on this exam
Each RBT practice test has 25 original questions with the answers explained. No account needed.
Open the RBT practice testsOriginal lesson written for RBT practice · updated October 25, 2026. Exam names are trademarks of their owners; tests.global is not affiliated with them.