What is ABA data collection?

In ABA, decisions are made from data. Throughout a session, technicians record how a client responds to each teaching trial, which prompts were needed, and any behaviors of interest, so progress can be measured objectively rather than by impression.

That data is then turned into graphs and reports a BCBA reviews to decide whether to advance a goal, change a strategy, or adjust the plan. Accurate, real-time capture keeps those decisions grounded in what actually happened.

Practically, data is often collected on phones or tablets in settings without reliable internet, so offline-capable capture that syncs later helps make sure nothing is lost or left for end-of-day paperwork.

Different goals call for different measures. Trial-by-trial data records whether each teaching opportunity was correct, prompted, or incorrect. Frequency and rate count how often a behavior occurs. Duration and latency time how long it lasts or how long it takes to start. Interval recording samples whether a behavior occurred during set blocks of time, which is more practical for behaviors that are continuous or very frequent.

The measure chosen has to fit the question being asked. Counting instances of a behavior that happens almost constantly produces a number nobody can act on, while a duration measure on a brief, discrete response says very little. Choosing badly here is one of the more common ways a program ends up with data that looks rigorous but cannot support a decision.

Data quality also depends on people collecting it the same way. Teams define behaviors in observable, unambiguous terms, train everyone to the same definitions, and periodically check agreement between observers. Combined with consistent capture during the session rather than reconstruction afterwards, that is what makes the resulting graphs trustworthy enough to change a treatment plan on.