Mount Etna Global Behaviour Academy
Clinical training
BCBA Test Content Outline · 6th edition

Visual Task List & Training Guide

Concept graphics for supervision, fieldwork and study. Every one of the 104 tasks across the nine domains, drawn to a single visual system, with a worked clinical example on each page.
Mount Etna Global Behaviour AcademyClinician

The nine domains

BCBA TCO 6th ed.
ABehaviorism and philosophical foundations5 tasks · 8 questions
BConcepts and principles24 tasks · 24 questions
CMeasurement, data display and interpretation12 tasks · 21 questions
DExperimental design9 tasks · 13 questions
EEthical and professional issues12 tasks · 22 questions
FBehavior assessment8 tasks · 23 questions
GBehavior-change procedures19 tasks · 25 questions
HSelecting and implementing interventions8 tasks · 20 questions
IPersonnel supervision and management7 tasks · 19 questions
104Tasks in total175 scored questions on the examination
Eight teaching cards first, then one graphic per concept, in domain order.
Part one

Concept cards

The eight teaching cards from the front of the guide: naturalistic teaching, language support, prompting and consequence systems.
Naturalistic teaching

Decision tree: least to most support

ESDM
1Wait for the responseSet up the activity, pause, and give the child room to act on their own.
2PromptAdd the lightest prompt that will produce the response.
3Change the materialsSwap in a higher preference item before you add more support.
4Change the promptMove up the hierarchy, then fade back down as soon as the child succeeds.
Functional spontaneous behaviour first
The goal of every step is a response the child produces for a natural reason, in a real activity.
ExampleBubbles are closed on the table. Wait 5 s, hold them up, then swap for the wand, then model “open”.
Move up one step at a time, and step back down the moment the child responds.
Naturalistic teaching

Joint attention framework

ESDM
01
Orient the response
Get the child looking towards you before you say anything.
ExampleHold the car at eye level and wait for Ava to look up.
02
Introduce the object
Bring the toy into the shared space between you both.
ExamplePlace the car on the mat directly between you.
03
Dyadic engagement
Back and forth turns with the object between you.
ExampleRoll the car to Ava, she rolls it back, eight times.
04
Support engagement
Add words and repetition to hold the episode longer.
ExampleSay “car” on every roll, then “fast car” once the turns are steady.
Each step builds on the one before, inside an activity the child chose.
Language support

The one up rule

Modelling
Add one or two words to whatever the child said
Your model sits one step above their current level, so it stays reachable.
One up1 word2 words3 words
Child“Car”“Drive car”
Child“Up”“Pick me up”
Child“More juice”“I want more juice”
Jumping three or four words above the child leaves the model out of reach, so keep the step small.
Model the next step up, and give access to the item either way.
Language support

Gestalt language processing

Echoic scripts
What it looks like
Whole chunks of language are learned and used as single units
“Where is the toilet?” “Où sont les toilettes?” “Time to go now”
The learner has the whole phrase before the individual words, in the same way a traveller learns one useful sentence in a new language.
“Where is the toilet?” is used for every request, not only for the bathroom.
1. Accept the script
Respond to the intention behind the phrase rather than correcting the form.
2. Model a new chunk
Give a phrase that fits the moment: “I need the toilet”, “I want a turn”.
3. Mix and match
As the chunks build, the learner starts recombining parts into flexible phrases.
Record which scripts the learner already has, and pick the next chunk from the activities they choose most.
Meaning first, grammar later.
Naturalistic teaching

The spotlight

ESDM
Shared focus
Adult
The activity the child chose
Child
Both people attend to the same object or action at the same time. That overlap is the spotlight.
The adult sits where the child is already looking
Position yourself face to face, at the child's eye level, with the materials between you. Your face and the toy share the same spotlight, so attending to the activity means attending to you.
Face to face
Sit opposite, not beside or behind.
Eye level
Drop to the floor or use a low chair.
Materials in front
Keep the toy in the shared space.
One activity
Clear the table of competing items.
ExampleBlowing bubbles while seated on the floor opposite the child, holding the pot at chin height.
Where you sit is a teaching decision.
Behavior-change procedures

Token economies and response cost

G.4
Token economy
Tokens are earned for target behaviour and exchanged for a backup reinforcer. Keep the exchange rate reachable.
1. Define the behaviour that earns a token
2. Set the exchange rate and the menu
3. Deliver tokens immediately, with praise
4. Thin the rate as the skill becomes fluent
ExampleFive tokens for completed maths items buys ten minutes on the scooter.
Response cost
Earned tokens are removed after the target behaviour. This is negative punishment, so it needs a strong earning system underneath it.
Never let the balance fall below zero
Keep earning opportunities far more frequent than losses
Pair every removal with a reminder of how to earn back
ExampleOne token is removed for grabbing, with two more available in the next five minutes.
Build the earning side first, and use cost sparingly.
Interaction strategies

ROCK: it takes two to talk

Hanen
R
Repeat what you say and do
Same words, same actions, so the pattern becomes predictable.
“Ready, set, go!” every single push of the swing.
O
Offer a chance to take a turn
Pause, look expectant, and wait long enough for a response.
Hold the bubble wand still and count silently to five.
C
Cue the child to take their turn
Add a gesture, a look, or a partial word to prompt the turn.
Lean in and say “ready, set…”.
K
Keep it fun, keep it going
Stay in the activity while the child is enjoying it, then stop before they do.
Eight swing pushes with a turn each time.
Two people, taking turns, in an activity the child chose.
Prompting

Say, show, do

Least to most
1
Say
Give the instruction once, clearly, then wait five seconds.
Example“Get your shoes.”
2
Show
Model the response, or point to what the learner needs.
ExamplePick up your own shoes while they watch.
3
Do
Provide physical assistance, then fade the contact on the next trial.
ExampleGuide the hand to the shoes, then release.
Then fade
Record the level used on every trial, and aim for the level below it next time.
Stop at the first level that works.
A
Domain A

Behaviorism and philosophical foundations

Five tasks, 8 exam questions. The goals of the science, its philosophical assumptions, and what separates the four levels of the field.
Behaviorism & philosophical foundations

The three goals of behaviour analysis as a science

A.1
Goal 01
Description
Systematic observation, recorded in a way that can be quantified and classified. It produces the facts the other two goals depend on.
Goal 02
Prediction
Repeated observation reveals correlation between events, so one can be used to anticipate the other. Correlation is not yet a cause.
Goal 03
Control
A functional relation is demonstrated: changing one variable reliably changes the behaviour. This is what makes intervention possible.
Description supports prediction, and only control demonstrates a functional relation.
Behaviorism & philosophical foundations

Five philosophical assumptions of the science

A.2
01SelectionismBehaviour is selected by its consequences, in the same way traits are selected in evolution.
02DeterminismThe universe is lawful and orderly, so events have causes that can be discovered.
03EmpiricismClaims rest on objective observation and measurement, not on opinion or authority.
04ParsimonySimple, logical explanations are ruled out before more complex ones are considered.
05PragmatismA concept is judged by whether it works, and by whether it produces useful outcomes.
These assumptions are what make behaviour analysis a natural science rather than a set of techniques.
Behaviorism & philosophical foundations

Mentalism, and how to avoid it

A.3
Hypothetical construct
An unobserved process or entity that is assumed to be present.
Example“Conscience.”
Explanatory fiction
A fictional variable used to explain the behaviour it is meant to describe.
Example“Tim is well behaved: he follows his conscience.”
Circular reasoning
Faulty logic in which the effect that is observed is mistaken for the cause.
Example“Tim is well behaved because he follows his conscience, and I know he follows his conscience because he is well behaved.”
Instead write“Tim completed 9 of 10 tasks without a prompt”, which is observable, measurable, and open to intervention.
Describe what the person did and what followed, not what they supposedly have.
Behaviorism & philosophical foundations

The four domains of behaviour analysis

A.4
01
Philosophy and conceptual analysis
Radical behaviourism is the underlying philosophy, and it informs the other three domains. Public and private behaviour follow the same principles.
02
Experimental analysis of behaviour
Theory-driven research on basic processes, run in laboratory or contrived settings and across species, with tight experimental control. It discovers the principles.
03
Applied behaviour analysis
Applied research in the natural environment, mostly with people, aligned to the seven dimensions. It develops the technologies practitioners use.
04
Behaviour service delivery
Professional practice: using those technologies to meet a client's immediate needs, with socially valid methods and results. This is the work we do in clinic and in homes.
Moore & Cooper (2003): the domains inform each other from philosophy through to service.
Behaviorism & philosophical foundations

The seven dimensions of applied behaviour analysis

A.5
01AppliedBehaviour of social significance: dressing, asking for help, crossing a road safely.
02Behavioural“Hits with an open hand”, rather than “is frustrated”: observable and countable.
03AnalyticA reversal design shows the token board, and only the token board, raised on-task time.
04TechnologicalThe protocol names the prompt, the delay in seconds and the reinforcer, so anyone can run it.
05Conceptually systematicThe plan is described as differential reinforcement, tied to the principles it comes from.
06EffectiveTantrums fell from 40 to 4 per week, and the family now eats out together.
07GeneralityThe skill holds six months later, at home, at school and with the grandparents.
Baer, Wolf & Risley (1968): the standard a programme is judged against.
B
Domain B

Concepts and principles

Twenty-four tasks, 24 exam questions. The largest concept domain: contingencies, schedules, stimulus control, verbal behaviour and response allocation.
Concepts & principles

Behaviour, response and response class

B.1
The widest term
Behaviour
Anything a living organism does that involves action and interaction with the environment. The whole category.
One instance
Response
One single instance of behaviour. A specific occurrence, at a specific time, that can be counted.
Same function
Response class
A group of responses that differ in form but produce the same effect on the environment, so they serve the same function.
Programmes target response classes, and data are taken on individual responses.
Concepts & principles

Stimulus and stimulus class

B.2
Single event
Stimulus
An energy change that affects the organism through its receptors. One specific event the learner can detect.
Grouped by shared property
Stimulus class
A group of stimuli that share a formal, temporal or functional property, and that occasion the same response.
Teaching to a stimulus class is what makes a skill usable outside the teaching setting.
Concepts & principles

Respondent conditioning

B.3
Before conditioningUSloud fire alarmURstartle, hands over ears
During conditioningNS + USassembly bell, then alarmURstartleRepeated pairing, with the bell presented just before the alarm.
After conditioningCSassembly bell aloneCRstartle, hands over ears
Reflexive relations are elicited by antecedents. Operant relations are strengthened by consequences, which is why the same behaviour can need two different plans.
Pairing builds the relation, and repeated unpaired presentations extinguish it.
Concepts & principles

Reinforcement and punishment

B.4 & B.5
Behaviour increases
Behaviour decreases
+Stimulus added
Positive reinforcement
Something is added after the response and the response becomes more likely.
ExamplePraise and a sticker follow hand raising, and hand raising increases.
Positive punishment
Something is added after the response and the response becomes less likely.
ExampleAn extra tidying task follows throwing, and throwing drops.
−Stimulus removed
Negative reinforcement
Something is removed after the response and the response becomes more likely.
ExampleSaying “break please” removes the demand, and the request increases.
Negative punishment
Something is removed after the response and the response becomes less likely.
ExampleTwo tokens are removed after grabbing, and grabbing drops.
Extinction sits outside the grid: the response stops producing its reinforcer at all.
Concepts & principles

Decoding a scenario in three questions

B.4 & B.5
1
Which behaviour changed?
Read who the question is asking about. That person's behaviour is the one with a consequence and a future frequency.
2
What was the consequence?
Something was added, or something was removed. Added is positive, removed is negative. Nothing to do with good or bad.
3
Did the behaviour increase or decrease?
Increase means reinforcement, decrease means punishment. Read it from what happened next time.
Added
Positive
Removed
Negative
Increased
Reinforcement
Decreased
Punishment
Worked scenario
A staff member hands over the iPad when Sam screams. Sam screams more often at the next tidy-up.
1. Sam's screaming.
2. The iPad was added.
3. Screaming increased.
Positive reinforcement of screaming
Same three questions every time, in the same order.
Concepts & principles

Automatic and socially mediated contingencies

B.6
No mediator
Automatic
The consequence follows directly from the behaviour, with no other person involved. Example: scratching an itch relieves it.
Another person delivers it
Socially mediated
Another person delivers the consequence. Example: asking for a snack, and a staff member hands it over.
This distinction decides whether an intervention can work through the people in the room.
Concepts & principles

Unconditioned, conditioned and generalised

B.7 & B.8
Unconditioned
Conditioned
Generalised
Reinforcers
Effective without any learning history, because of phylogeny. Example: food, water, warmth.
Became reinforcing by being paired with an existing reinforcer. Example: praise, a favourite song.
Paired with many different reinforcers, so it works under many conditions. Example: tokens, money.
Punishers
Suppress behaviour without learning history. Example: intense heat, a painful impact.
Became punishing through pairing. Example: the word “no”, a particular tone of voice.
Paired with many punishers, so it suppresses across conditions. Example: losing tokens already earned.
Generalised means paired with many, not stronger. Effect is always defined by what happens to the behaviour.
Concepts & principles

Basic schedules of reinforcement

B.9
Ratio: counted responses
Interval: elapsed time
Fixed
FRevery fifth response
Steady high rate, with a pause straight after each reinforcer.
ExampleFR 5: a token after every five completed maths items.
FIfirst response after 2 min
Responding builds as the end of the interval approaches.
ExampleFI 2 min: the first hand raise after two minutes earns attention.
Variable
VRon average every fifth
High steady rate, and the most resistant to extinction.
ExampleVR 5: praise after 3, then 7, then 5 responses.
VIon average every 2 min
Low steady rate that persists over long sessions.
ExampleVI 2 min: the teacher checks in at 90 s, then 3 min, then 2 min.
Post-reinforcement pause
Fixed schedules are predictable, so responding pauses after each reinforcer.
Limited hold
A set time to respond once the interval ends, before the next one begins.
Thinning and thickening
Thinning raises the ratio or interval; thickening lowers it.
Reinforce densely to teach a skill, then thin to a variable schedule to maintain it.
Concepts & principles

Seven compound schedules of reinforcement

B.10
Read each one on two featuresHow many response classes?Is an SD signalling which schedule is in effect?
Concurrent
2+ classesSignalled
Two or more behaviours on independent schedules, available at the same time. This is where matching law applies.
Multiple
1 classSignalled
Independent and successive schedules, presented in random order, each with its own SD.
Mixed
1 classNo SD
Exactly like a multiple schedule, except no SD tells the learner which schedule is running.
Chained
1+ classSignalled
Independent and successive schedules that must be completed in a specified order.
Tandem
1+ classNo SD
A specified order like a chained schedule, but with no SD marking the components.
Alternative
1 classRatio + interval
A ratio and an interval schedule run together, and reinforcement follows whichever is met first.
Conjunctive
1 classRatio + interval
Also a ratio and an interval together, but both must be met before reinforcement is delivered.
Quick test
Alternative is “either”. Conjunctive is “both”. Mixed and tandem are the unsignalled twins of multiple and chained.
Cooper, Heron & Heward (2012): compound schedules combine the basic schedules with DRO variants and extinction.
Concepts & principles

Behaviour contrast

B.10
Responding changes in one unchanged component of a multiple schedule, because something changed in another component.
Positive contrast
Intervention decreases the behaviour in one setting, and it increases in another
School: DRA in place, shouting falls
Home: nothing changed, shouting rises
Nothing was done at home, so the rise there is contrast, not a new problem.
Negative contrast
Intervention increases the behaviour in one setting, and it decreases in another
Clinic: requesting is reinforced richly, requests rise
School: unchanged, requesting falls
Check the other settings before you conclude a programme has failed or succeeded.
Mayer, Sulzer-Azaroff & Wallace (2014): take data in every setting the behaviour occurs in.
Concepts & principles

Extinction: what to expect

B.11
The pattern on the graph
BaselineBurst and recoveryExtinction
The burstWhen reinforcement stops, intensity, rate and variability of the behaviour usually increase first.
Spontaneous recoveryA previously extinguished behaviour reappears even though it has not been reinforced. Stay calm and keep the plan in place.
ResurgenceAn older behaviour returns when the behaviour that replaced it goes on extinction.
Before you implement it, check
Have earlier extinction attempts failed and taught the learner that persistence pays? Was the behaviour on an intermittent schedule? Does it take low effort to produce? Any of these predicts resistance.
Extinction is never used on its own here: pair it with a reinforced replacement behaviour that gets the same outcome.
Mayer, Sulzer-Azaroff & Wallace (2014): plan for the burst before it happens.
Concepts & principles

Stimulus control and discrimination

B.12 & B.13
B.12
Stimulus control
A behaviour occurs more often in the presence of a particular stimulus, because that stimulus has been correlated with reinforcement for it.
SD presentresponding is likely
SΔ presentresponding is unlikely
ExampleAsking for a snack when a staff member is in the kitchen, but not when the room is empty.
B.13
Stimulus discrimination
Responding differs across two or more stimuli, because reinforcement has been available in the presence of one and not the other.
SD
ExamplePressing the green button, which opens the door, and not the red one, which does nothing.
Discrimination training builds the control; the goal is responding under the right conditions, not more responding.
Concepts & principles

Stimulus and response generalisation

B.14
Stimulus generalisation
Novel stimuli evoke the trained response
Stimulus 1Stimulus 2Stimulus 3Stimulus 4
One response
ExampleTrained to wash hands at the clinic sink, Ana washes her hands at the school sink and at home.
Response generalisation
A trained stimulus evokes novel responses
One stimulus
Response 1Response 2Response 3Response 4
ExampleTaught to ask “can I have a turn?”, Ana also says “my turn now” and “can I play?”.
Baer, Wolf & Risley: generalisation should be programmed, rather than expected or lamented.
Concepts & principles

Motivating operations, and how they differ from stimulus control

B.16 & B.17
Establishing operation
Value goes up, behaviour goes up
Deprivation raises the value of the reinforcer, and behaviour that has produced it becomes more frequent. Example: a missed lunch makes snack requests more likely.
Abolishing operation
Value goes down, behaviour goes down
Satiation lowers the value of the reinforcer, and behaviour that produced it becomes less frequent. Example: after free access to the iPad, requesting it drops.
Telling them apart
MOAlters how much the consequence is worth right now.
SDSignals whether the consequence is available right now.
A thirsty child walking past a water fountain: thirst is the MO, and the fountain is the SD. Remove the thirst and the fountain no longer occasions the behaviour.
Check the MO before every teaching trial: no motivation, no learning opportunity.
Cooper, Heron & Heward: MOs are value-altering, discriminative stimuli are availability-signalling.
Concepts & principles

Rule-governed and contingency-shaped behaviour

B.18
Learned by doing
Contingency-shaped
The behaviour was shaped by direct contact with its consequences, over repeated exposure. It is usually smooth, fast and hard to describe.
ExampleA staff member who has run hundreds of sessions adjusts pacing without thinking about it.
Learned by being told
Rule-governed
The behaviour is controlled by a verbal statement of the contingency, so the learner does not have to contact the consequence first.
ExampleA new employee follows the written protocol on their first day and gets it right.
Rules get behaviour started quickly; contact with the actual consequences is what makes it fluent.
Concepts & principles

Six elementary verbal operants

B.19
Antecedent controlThe operant it produces
Motivating operationMandA request for a reinforcer. Example: “bubbles” when the bubbles are out of reach.
Non-verbal stimulusTactNaming or identifying something present. Example: seeing a dog and saying “dog”.
Verbal stimulusIntraverbalAnswering questions and holding conversation. Example: “How was school?” “We did painting.”
Vocal stimulusEchoicRepeating what was heard, point to point. Example: “say ball”, “ball”.
Written stimulusTextualReading written words aloud. Example: reading “stop” from a sign.
Vocal stimulusTranscriptionWriting what is heard. Example: writing a spelling word that was read out.
Skinner (1957): verbal behaviour is reinforced through the mediation of another person.
Concepts & principles

Multiple control in verbal behaviour

B.20
Many variables, one response
Convergent multiple control
A single response is controlled by more than one antecedent variable at the same time.
Sees a ballHears “what is it?”Wants a turn“ball”
One variable, many responses
Divergent multiple control
A single antecedent variable evokes more than one different response.
Sees a dog“dog”“animal”“that is Rex”
Most everyday utterances are multiply controlled, which is why single-variable teaching often fails to generalise.
Concepts & principles

Stimulus equivalence

B.21
Equivalence is accurate responding to untrained, non-reinforced stimulus relations, which emerges after some other relations have been reinforced.
ReflexivityA = AGeneralised identity matching, and a prerequisite for the rest.
SymmetryA = B, so B = AAfter one direction is trained, the reverse direction is derived.
TransitivityA = B, B = C, so A = CTwo trained relations produce a third that was never taught.
Worked example
.51/250%
Taught: .5 = 1/2 and .5 = 50%.
Derived: 1/2 = .5, 50% = .5, 1/2 = 50% and 50% = 1/2.
Teach two, get four free
Two taught relations produced four more without further teaching, and the three stimuli now form one equivalence class.
Plan teaching to exploit derived relations, then probe them before scoring mastery.
Concepts & principles

Behavioural cusps and pivotal behaviours

B.21
Behavioural cusp
Opens access to a new environment
The behaviour gives immediate access to new environments, stimuli and contingencies, and it competes with less appropriate behaviour.
DrivingEverywhere you can now go
In practiceIndependent toileting, or asking for help: one skill and the day changes shape around it.
Pivotal behaviour
Produces changes in untrained responses
Teaching it produces collateral change in behaviours that were never taught, and those untrained responses create new reinforcement opportunities.
Joint attention is established
The adult points to a card and says “find the same”
The learner points to the match in the array
Reinforcer delivered, and matching now generalises
A single behaviour can meet the criteria for both a cusp and a pivotal behaviour.
Rosales-Ruiz & Baer (1997): choose targets that unlock the most, not the easiest to measure.
Concepts & principles

Behavioural momentum and response persistence

B.22
Momentum describes how strongly behaviour persists when conditions change. A behaviour with a rich reinforcement history keeps going, in the same way a heavy object keeps moving.
Two components
Mass, or resistance to change
Built by rate, magnitude and quality of reinforcement over time.
Velocity, or current rate
How often the behaviour is happening right now.
Why it matters in session
A problem behaviour with years of dense reinforcement has high mass, so it will persist through the first weeks of a new plan.
High massLow mass
So build the replacement behaviour's mass deliberately: reinforce it densely, immediately and generously before thinning.
Momentum predicts persistence, so plan for it rather than reading early persistence as failure.
Concepts & principles

Matching law

B.23
On concurrent schedules, responses are allocated in line with the rate of reinforcement each option produces.
Two options, available at once
Response A6 reinforcers per minute
Response B2 reinforcers per minute
75%25%
Responding is distributed A to B on a 3:1 ratio, matching the reinforcement available.
The published example
Vollmer & Bourret (2000) applied matching law to college basketball players, and the allocation of two-point and three-point shots tracked the rate of reinforcement each shot produced.
Why it matters in session
Problem behaviour and a replacement behaviour are usually two options on a concurrent schedule. If the problem behaviour pays more often, faster or bigger, that is where responding will go.
So make the replacement behaviour the better deal: richer, quicker and less effort than the behaviour you are trying to reduce.
Rate, immediacy, magnitude and effort all decide which option wins.
Concepts & principles

Imitation and observational learning

B.24
Matched to a model
Imitation
The model is the antecedent, the response matches its form, and the match happens immediately after. Formal similarity and immediacy are both required.
ExampleA staff member claps, and the child claps straight after.
Learned from someone else’s consequence
Observational learning
New behaviour is acquired by watching someone else contact the consequence, without the observer being reinforced directly at the time.
ExampleA child sees a peer earn a token for tidying up, and starts tidying up at the next opportunity.
Imitation copies the form; observational learning acquires the contingency.
C
Domain C

Measurement, data display and interpretation

Twelve tasks, 21 exam questions. Defining behaviour, choosing a measure, checking it is trustworthy, and reading the graph.
Measurement & data

Writing an operational definition

C.1
Three tests it must pass
Objective — only observable events and their properties.
Clear — unambiguous, so it can be read and repeated.
Complete — boundaries stated, including what does not count.
Two observers reading the same definition should score the same event the same way, without asking you what you meant.
Not a definition
“Sam gets frustrated during maths.”
Frustration is a private state, so no two observers will agree on when it started or stopped.
A definition
“Sam pushes materials more than 15 cm across the desk, or leaves his seat without asking.”
Both parts are observable, countable, and have a clear start and stop.
Write the definition before the first data point, and check it with a second observer.
Measurement & data

Direct, indirect and product measures

C.2
Watch it happen
Direct
The behaviour itself is observed and recorded as it happens, by someone present.
ExampleTallying hand raises during a lesson.
Ask about it
Indirect
Information comes from someone’s report rather than observation, so it is filtered by memory and interpretation.
ExampleA caregiver interview or a behaviour rating scale.
Measure what it left behind
Product
A permanent result of the behaviour is measured after the fact, rather than the behaviour itself.
ExampleCounting completed maths items, or photographing a tidied room.
Direct measures are the strongest; indirect measures generate hypotheses rather than confirm them.
Measurement & data

Dimensional qualities of measurement

C.3 & C.4
01Repeatability
The response can occur again and again, so you can count it.
CountRateCeleration
Example
Mia raised her hand 12 times in a 30 minute lesson: a rate of 0.4 per minute.
02Temporal extent
The response occupies time, so you can measure how long it lasts.
Total durationDuration per occurrence
Example
Jonah stayed in his seat for 14 of the 20 minute work block, in three separate stretches.
03Temporal locus
The response occurs at a point in time, so you can measure when.
LatencyInter-response time
Example
After the instruction “get your shoes”, Ava started moving 9 seconds later: a latency of 9 s.
Each dimension answers a different question: how many, how long, when.
Measurement & data

Continuous vs. discontinuous measurement

C.5 & C.6
ContinuousEvery instance
Every occurrence in the observation period appears in the data.
Event recording
Frequency · rate · percentage · trials to criterion
Timing
Duration · latency · inter-response time
Example
23 hand flaps counted across the full 15 minute play session.
DiscontinuousSampled
Only some intervals are scored, which builds a known bias into the data.
Partial intervalOverestimates
Whole intervalUnderestimates
Momentary time samplingApproximates
Example
Partial interval scores 8 of 15 intervals on-task (53%), though Sam was on task 4 minutes.
Sample when continuous recording is impractical, then state the bias it introduces.
Measurement & data

Derivative measures

C.7
Data you already collectedA derivative measure is calculated from data on hand, so it costs no extra observation time.
01Percent of occurrence
8÷10=80%
Responses divided by opportunities. Bounded 0 to 100%, so it conceals the number of opportunities. Always report the denominator.
Example
Leo handed over the picture card on 8 of 10 requesting opportunities during snack: 80% independent mands.
02Trials to criterion
14 trials to three consecutive correct
Opportunities needed to reach a set performance level. The cleanest way to compare two teaching procedures.
Example
Errorless prompting taught tooth brushing in 14 trials; least-to-most prompting took 31.
Calculated from data on hand, and always reported alongside the raw counts.
Measurement & data

Interobserver agreement

C.8
MeasureHow agreement is calculatedWorked example
Event recordingTotal count, mean count-per-interval, exact count-per-interval, or trial-by-trial.92%Observer A counted 23, observer B 25: 23 ÷ 25.
TimingTotal duration, or mean duration per occurrence, compared between observers.95%190 s and 200 s of crying recorded.
Interval methodsInterval-by-interval, scored-interval, or unscored-interval agreement.87%Both observers agreed on 13 of 15 intervals.
20%Collect IOA on at least a fifth of sessions, across observers and conditions, and choose the calculation before you look at the data.
Agreement measures consistency between observers, and accuracy is a separate question.
Measurement & data

Trustworthy measurement

C.8
01
Accuracy
The observed value matches the true value of the event.
ExampleA video recount confirms the 23 hand flaps scored live.
02
Reliability
Repeated measurement of the same event yields the same value.
ExampleThree RBTs score the same session tape within one count of each other.
03
Validity
The data are directly relevant to the question being asked.
ExampleMands per hour answers a communication goal; seat time does not.
Believable data
All three qualities hold at once, and the clinical decision rests on them.
Check all three before a graph reaches a parent meeting.
Measurement & data

Measurement artifacts

C.9
Data that misrepresent what actually happened. Three usual suspects, each one a flaw in how the system was designed.
01
Discontinuous measurement
Sampling misses instances, or credits a whole interval to one brief response.
Example
One 2 second shout scores the full 5 minute interval, so the graph reads 100%.
02
Poorly scheduled observations
Observing when the behaviour is unlikely, or only when it is at its worst.
Example
Aggression data collected only during preferred free play shows a flat baseline.
03
Insensitive or unrepresentative scales
A scale too coarse to show real change, or a ceiling the learner has already reached.
Example
A 3 point rating hides a drop from 40 to 12 tantrums per week.
Ask of every graph: could the method itself have produced this shape?
Measurement & data

Graphing and interpreting data

C.10 & C.11
Choosing the display
Equal-interval line graph — the default. Shows level, trend and variability over time.
Bar graph — for comparing discrete conditions with no time series.
Cumulative record — total responses only ever rise; slope shows rate.
Reading it: four questions
Level
Where on the y-axis is the data sitting within this phase?
Trend
Which direction is it moving, and how steeply?
Variability
How much do points bounce around that trend?
Change across phases
How quickly did it change at the phase line, and do the phases overlap?
Label the axes, mark every phase change, and never join data points across a phase line.
Measurement & data

Precision teaching

C.10 & C.11
Ogden LindsleyA measurement-led approach: pinpoint the behaviour, chart it daily, and let the data decide what changes.
Pinpoints
The target is directly observable, measurable and active. Something the learner does, not something they are.
Learning channels
The input and output pair is named: see-say, hear-write, see-do.
Standard celeration chart
A semi-logarithmic chart, so rate changes are read the same way by everyone.
Fluency
Accuracy plus speed. Fluent performance brings retention, endurance, stability and application.
Aims
A stated rate range that counts as fluent for this behaviour, set before teaching starts.
The learner is always right
If the data are not moving, the programme changes, not the learner. Decisions come from the chart.
Binder & Watkins (1990): pick a target the learner can actively do, and chart it every day.
Measurement & data

Measuring procedural integrity

C.12
Procedural integrity is the extent to which the plan is delivered as written. Without it, a flat graph cannot tell you whether the intervention failed or was never run.
Dimensions to measure
Accuracy — were the steps performed correctly?
Dosage — how many opportunities were delivered, out of how many planned?
Consistency — was it the same across staff, settings and days?
A step-by-step integrity checklist
Materials set out before the session
SD delivered once, as scripted
Prompt delivered at the planned delay
Reinforcer delivered within 3 seconds
Data recorded before the next trial
80%Four of five steps correct. Retrain the prompt-delay step before drawing any conclusion from the client’s data.
Take integrity data on the same schedule as client data, not only when something looks wrong.
D
Domain D

Experimental design

Nine tasks, 13 exam questions. Variables, validity, the single-case designs, and the analyses that answer a specific question.
Experimental design

Dependent and independent variables

D.1
The cause you control
Independent variable
What you manipulate. The intervention, the schedule, the prompt level. It is under the practitioner’s control.
ExampleIntroducing a 4 minute FT reinforcement schedule.
The effect you measure
Dependent variable
What you measure. The behaviour of interest, which is expected to change because the independent variable changed.
ExampleRate of shouting per hour.
Functional relationDemonstrated when the dependent variable changes reliably with the independent variable, and with nothing else. That is the whole purpose of the design.
Only one independent variable changes at a phase line, or the relation cannot be attributed.
Experimental design

Three kinds of validity

D.2
IV
to DV
Internal
The change in the dependent variable can be attributed to the independent variable alone.
Example
On-task time rises when the token board goes in, drops when it is withdrawn, and rises again on return.
Who
else?
External
The finding generalises to other people, settings, behaviours and times.
Example
The same token procedure works for four other learners in two different classrooms.
Worth
it?
Social
The goals, the procedures and the outcomes are acceptable to the client and the family.
Example
Dad rates the plan 5 of 5 for fit with the family routine at bedtime.
A confounding variable moves with the IV, and it is the reason a strong effect can still mislead.
Experimental design

Threats to internal validity

D.3
History
An event outside the study occurs between phases and could explain the change.
ExampleA medication change starts the same week as the intervention.
Maturation
The participant changes over time through development or fatigue, independent of the intervention.
ExampleA four-year-old’s language grows over a six-month baseline.
Instrumentation
The measurement itself changes: a new observer, a revised definition, a different recording sheet.
ExampleA second observer scores partial interval more loosely.
Attrition
Participants drop out, so the remaining data no longer represent the original group.
ExampleTwo of four children leave the programme mid-study.
Testing and reactivity
Repeated measurement itself changes the behaviour being measured.
ExampleStaff perform differently while being observed.
Regression to the mean
An extreme baseline drifts back toward typical levels on its own.
ExampleIntervention starts on the worst week of the term.
Internal validity is about ruling out rival explanations, and design is how you rule them out.
Experimental design

Defining features of single-case designs

D.4
01Individuals as their own controlsEach participant is compared against their own performance under a different condition, not against a group average.
02Repeated measuresThe dependent variable is measured many times per condition, so trend and variability are visible.
03PredictionA stable baseline lets you state what would happen next if nothing changed.
04VerificationReturning to the prior condition shows the predicted pattern really was under baseline control.
05ReplicationThe effect is reproduced at least once more, so it is not a one-off.
Prediction, verification and replication are the three-part logic every single-case design rests on.
Experimental design

Single-case and group designs

D.5
Single-case designs
Strengths
Answers what works for this client
Ongoing data allow the plan to change mid-course
Works with the small numbers real practice provides
The individual’s pattern is never averaged away
Weaker on generality: results have to be replicated across people to extend beyond the case.
Group designs
Strengths
Supports generality across a population
Random assignment controls many threats at once
Comparison across treatments at scale
Statistical estimates of effect size
Weaker for practice: an average effect does not tell you what to do for the person in front of you.
Group data inform what to try; single-case data tell you whether it worked for this client.
Experimental design

Four single-case experimental designs

D.7
A-B-A-B
Reversal
Conditions alternate between baseline and intervention. Withdrawing the intervention should return responding to baseline, which verifies the effect.
Staggered starts
Multiple baseline
The intervention is introduced at staggered times across behaviours, settings or participants. No reversal is needed, so it suits irreversible skills.
Rapid alternation
Multielement
Two or more conditions alternate rapidly, each signalled by its own stimulus, so their effects can be compared side by side. This is the functional-analysis design.
Stepped criteria
Changing criterion
One behaviour, and the criterion for reinforcement is raised in planned steps. The effect is shown when responding tracks each new criterion.
Pick the design by whether the behaviour can reverse, and by how many conditions you need to compare.
Experimental design

Three ways to analyse a treatment

D.8
Parametric
The value of the independent variable is manipulated. The intervention stays the same; only its size changes.
FT 2 minute NCRFT 3 minute NCRFT 4 minute NCR
QuestionHow much of it do we need?
Nonparametric
The independent variable is either part of the phase or it is not. The intervention is simply on or off.
Visual schedule usedNo visual scheduleVisual schedule used
QuestionDoes it work at all?
Component
One component is removed from a treatment package at a time, to find out whether it is necessary.
DRA + extinction + tokensDRA + extinctionDRA
QuestionWhich parts are carrying the effect?
Name the analysis before you change a phase, so the graph can answer the question you asked.
E
Domain E

Ethical and professional issues

Twelve tasks, 22 exam questions. The core principles, confidentiality, boundaries, cultural humility and the requirements we work under.
Ethical & professional issues

The four core principles

E.1
Principle 01
Benefit others
Work to maximise benefit and do no harm, protect the welfare and rights of clients above all else.
ExampleChoosing the less intrusive procedure that the data support.
Principle 02
Treat others with compassion, dignity and respect
Consider each person’s needs and perspective, and behave toward them as you would want to be treated.
ExampleAssent is sought and honoured, not assumed.
Principle 03
Behave with integrity
Be truthful, follow through on obligations, and hold yourself to the same standard you hold others.
ExampleCorrecting your own error in a report before anyone finds it.
Principle 04
Ensure competence
Practise only within the boundaries of your training and experience, and keep that competence current.
ExampleDeclining a case type you have never been supervised in, and referring on.
BACB Ethics Code for Behavior Analysts: these four principles sit underneath every specific standard.
Ethical & professional issues

Who is harmed when conduct slips

E.2
Risk 01
Harm to the client
Ineffective or intrusive procedures, lost learning time, and broken trust fall directly on the person you serve.
ExampleContinuing a plan the data show is not working.
Risk 02
Harm to the profession
Each breach erodes public trust in behaviour analysis and makes every colleague’s work harder.
ExampleMisrepresenting credentials or outcomes to families.
Risk 03
Harm to yourself and society
Certification, employment and legal consequences for you; weaker protections and services for everyone.
ExampleAn ethics violation report that ends a certification.