Old system – try out Tx and then explain what happened New system – assess function prescribe Tx ABA 2: Functional Analysis.

• Old system – try out Tx and then explain what happened

• New system – assess function prescribe Tx

ABA 2: Functional Analysis

• FA models– AB – shows functional relation between EO and

behavior; less compelling as reinforcers are not varied; usually done in extinction

• Behavior occurs because perhaps in extinction or accidental reinforcement

• Task stimuli may signal availability of attention, thus, results may be hard to differentiate

– ABC - more compelling, as putative reinforcer is varied.


– Why AB models? • ABC may strengthen new classes, but this seems rare (but

see some studies with tangibles)– Also these reinforcers are often already delivered in the

environment for problem behavior.

– Important to gather descriptive evidence of possible relations between behavior and tangibles.

– Moreover, creating new operants may help alert caregivers that this could happen in new environments.

• AB assessments can be quick and easy

– Best practice? ABC analyses




• Criteria for inclusion of review:– Separate FA, not just in Tx– Problem behavior – behavioral excess that someone

complains about– Direct observation and measurement of problem

behavior (no ratings, questionnaires, etc)– At least 2 conditions in which a variable was

manipulated. Studies that relied just on descriptive analysis were excluded.

– AB model and ABC model


• Kinds of elements noted:– Brief assessment - <= 2 observations; – Full > 2 observations– Several elements noted:

• Antecedent variables

• Consequences delivered

• Data display

• Design

• Session durations

• Kind of FA

• Topographies

• Population and setting

• Functions



• Results:– Population –kids with disabilities

– No FA with adults without disabilities– Setting –inpatient facilities, then schools, other– Topographies –SIB (64%) agg (40%) then disruption– Models: AB and ABC (87%)– Condition types -

• Most involved assessments of attention and escape

• Some have tangibles

• 60% examined self stim (but are indirect)



– Control conditions – • Full FA used no demands, toys, and free attention. EO s are

eliminated.

• Brief FA – test condition serves as control for another test (no demands in attention, and no attention in demands) – problem is that it is hard to see self stim or multiply controlled

– Session duration: 5-15 minutes – Assessment duration (# of conditions) – most done

until stability obtained• But Northrup looked at a brief FA in 90 min outpatient eval

with 1 or 2 sessions some studies did this model



– Experimental design – • Multi-element, ABAB in studies looking at a single

variable. – good if rapidly alternating conditions produce interactions among conditions

• W/D and Reversal

• Pairwise – each test condition is alternated with a control condition in a multi-element - used to minimize the interaction among conditions



– Antecedent events: • Most common - variations of social-positive – Consider

alone, attention, play– No attention vs divided attention

• Variations of social-negative – many studies have just examined demand as a single condition

– Only one examined novelty, duration of instruction, and rate of presentation when behavior produced escape from tasks

– Need more studies that identify important variables in task escape


– Consequent events• Type/quality – Attention is often reprimands with touch

– Tone of the reprimand can be important (Drama!)

• Access to tangibles

• Escape from tasks

• A study showed that access to self restraint strengthened SIB

• Schedule – FR 1 in most; some had intermittent:

• Duration – Attention for 5-10s, tangibles for 30s, and escape for 30 s


– Reinforcer duration issue:• No reinforcer time subtracted

– 10 in 5 minutes for attention (6 sec) = 2/min

– 6 in 5 minutes for escape (30 sec) = 1.2/min

• Reinforcer time subtracted:– Attention = 10 in 4 minutes = 2.5/min

– Escape = 5 in 2 minutes = 2.5/min


• Outcomes of analyses– Escape maintained in 34%, attention maintained in

25%, 10 % tangible, automatic – 10%– Most aggression was escape; stereotypies were

automatic reinf– In general, though, function can’t be predicted by

topography


• Recommendations/Best Practice– Limit response topographies in the analysis– ABC model instead of AB– Use control/test pairs unless time is limited– Include natural stimuli when possible – but ensure

integrity– Take advantage of EO s in assessment

– Alone Attention– Control Tangible??

– Include stimuli to signal conditions– Brief sessions (e.g., 10 minutes) if sufficient rates


• Recommendations/Best Practice– Test for automatic reinforcement– Test for functional relations between tangibles and

behavior only if descriptive suggests it– Start with brief and simple, and then to more complex – Use descriptive as adjunct to do more complex analyses


• Undifferentiated results– Look for odd variables– Use natural stimuli in natural environment– Minimize topographies; graph them separately


• Undifferentiated results– Increase # of observations


Old system – try out Tx and then explain what happened New system – assess function prescribe Tx ABA 2: Functional Analysis.

Documents

functional analysis

conditions aba

descriptive analysis

duration attention

functional relations

analysis abc model

tx aba

task escape aba