RBT

Continuous measurement records every instance of the behavior during the whole observation period. It gives the most complete and accurate picture of behavior and is preferred whenever the observer can attend to the client the entire time. The four continuous procedures on the exam are frequency, duration, latency, and interresponse time.

Procedure	WhatIsRecorded	BestFor	Example
Frequency(event	Thenumberoftimesthebehavior	Discretebehaviorswithaclear	Handraises mands hits correct , , ,
recording)	occurs;convertedtoratewhensession	beginningandend offairly ,	answers
	lengthsdiffer	uniformlength
Duration	Howlongeachresponselasts from ,	Behaviorswherethelengthof	Tantrums timeontask timein , ,
	onsettooffset;totaldurationor	eachepisodematters	seat crying ,
	durationperoccurrence
Latency	Timefromtheendofaninstructionor	Complianceandresponsespeed	Secondsfrom"Lineup"untilthe
	cue(theSD)tothestartofthe		childstartswalking
	response
Interresponsetime	Timebetweentheendofoneresponse	Pacingofresponses	Secondsbetweenbitesoffood;
(IRT)	andthestartofthenext		timebetweenvocaloutbursts

Frequency and rate. Count is the raw number of responses. When every session is the same length, count can be compared directly. When sessions differ in length, convert count to rate (count ÷ time) so sessions can be compared fairly. The BACB and most ABA texts use "frequency" and "rate" interchangeably for count per unit of time.

Duration. Start timing when the behavior begins and stop when it ends, as defined by the operational definition's onset and offset criteria (for example, a tantrum ends after 10 consecutive seconds with no crying or screaming). Report either total duration (all episodes added together) or duration per occurrence (the length of each episode, often averaged).

Latency versus IRT. Both measure time before a response. Latency begins at a stimulus (an instruction, a bell, a question); IRT begins at a previous response. Short latencies are usually the goal for compliance, while an intervention to slow down eating aims to lengthen the IRT between bites.

Discontinuous measurement samples behavior instead of recording every instance. The observation period is divided into equal intervals, and each interval is scored as an occurrence or a nonoccurrence. The result is reported as the percentage of intervals in which behavior was scored, which is an estimate, not an exact count. Discontinuous methods are used when continuous recording is impractical: the behavior is very frequent or has no clear beginning and end, the RBT is also teaching, or several clients or behaviors must be observed at once.

Procedure	IntervalIsScoredWhen	TendsTo	BestFor
Partial -interval	Thebehavioroccursatanytimeduring	Overestimate	Behaviorstodecrease;brief frequent ,
recording(PIR)	theinterval evenonceandbriefly ,		behaviors(eg callingout . ., ,
			stereotypy)
Whole-interval	Thebehavioroccursthroughoutthe	Underestimate	Behaviorstoincreasethatshouldbe
recording(WIR)	entireinterval		sustained (e .g on-task behavior ., ,
			cooperativeplay)
Momentarytime	Thebehaviorisoccurringattheexact	Over- or	Long-lasting continuous behaviors; ,
sampling(MTS)	momenttheintervalends	underestimate	observingseveralclientsorbehaviors
			atonce

Why the errors run in these directions. Partial-interval recording scores a whole 10-second interval for a one-second response, so it inflates the estimate; that bias is conservative when you want to show a problem behavior decreasing. Whole-interval recording drops an interval for a one-second interruption, so it deflates the estimate; that bias is conservative when you want to show an appropriate behavior increasing. Momentary time sampling misses anything that happens between the sampling moments.

Permanent-product recording measures the lasting outcome of a behavior after it occurs rather than watching the behavior itself: completed worksheets, items assembled, words written, dishes washed, beds made, or even damaged items or wounds. It lets the RBT do other things during the session and allows accurate checking later. Its limitations are that the product must reliably reflect the target behavior (someone else might have completed the worksheet) and that it tells you nothing about how the product was produced.

Exam scenarios often describe the procedure instead of naming it. "Marked the interval if the student was out of seat at any point" is partial-interval; "only if the student remained seated for the full interval" is whole-interval; "looked up when the timer vibrated and recorded whether the student was out of seat at that moment" is momentary time sampling.