What is baseline data in ABA?
Baseline data is the measurement of a behavior or skill taken before intervention begins - the starting point every later claim of progress is measured against.
Before a program teaches anything, it establishes where the learner already is. Baseline data is that record: repeated measurement of the target under ordinary conditions, with no teaching procedure running, so that both the level of the behavior and how much it varies are known before anything changes them.
The point is comparison. Without a baseline, a graph showing a client tacting twenty items says nothing - twenty may be an improvement on two or a decline from thirty. Baseline is what turns a number into evidence, which is why it is collected for behavior reduction targets and skill acquisition targets alike.
A usable baseline is stable rather than merely brief. Clinicians generally continue baseline sessions until the data settle into a discernible pattern, because a single unrepresentative session followed immediately by intervention invites the wrong conclusion - behavior already trending on its own gets credited to the program. Measuring the same way in both phases matters just as much: a baseline taken by frequency and a treatment phase taken by interval recording are not comparable.
Baseline is also a documentation requirement. Treatment plans record it for each goal alongside the mastery criterion, and progress reports return to it at every reauthorization. That is the practical reason baseline belongs in the plan rather than in a clinician's private notes - a funder is being asked to pay for the distance between two numbers, and both of them have to be on the record.
Baselines recur throughout a case rather than happening once. New goals get their own as they are introduced, and probes of mastered targets re-establish where a skill stands after time has passed. A practice whose data system makes taking a baseline as routine as running a teaching trial ends up with programs that can actually be evaluated.