Behavior Assessment

How behavior is assessed, one task at a time: where the data come from, what each method can and cannot show, and how results turn into decisions.

Concepts
8
Sections
5
Diagrams
5
In this chapter
  1. 01 Before the first observation
  2. 02 Strengths and skill needs
  3. 03 Preferences and reinforcers
  4. 04 Finding the function of behavior
  5. 05 From data to decisions

Before the first observation

An assessment starts before anyone watches the behavior. What is already on file, and what matters to the family, decides which questions are worth asking and where to look for the answers.

F.1

Reviewing records at the start of a case

Records from school, medical providers and earlier services tell you what has been tried, what changed and what might be going on in the body. Read them to sharpen the referral question and to decide what to observe first. Pain, poor sleep, illness, medication changes and hearing or vision problems can all change behavior, so a possible medical contributor is flagged and passed to a qualified provider; the behavior analyst does not diagnose or treat it. Ask only for records that bear on the question, and only with the authorization required.

Example

A nine-year-old has started screaming during morning classes. His file shows a new allergy medication started three weeks ago, and that a previous provider thought the screaming was about attention. You ask the family to check the medication with the prescribing doctor, and you plan observations that can test the attention idea instead of taking it as given.

Common confusion

What a file says is a lead, not a finding. An earlier hypothesis, a past diagnosis of function or a note that "nothing worked" still has to be checked against current data. The opposite error also costs points: requesting every record you can find, relevant or not, or obtaining records without consent.

Key points (4)
  • At the outset, educational, medical, historical and prior-service records are reviewed to frame the assessment questions and plan what to observe.
  • Medical and physiological factors (pain, illness, sleep, medication effects, sensory impairments) can affect behavior; possible contributors are referred to and coordinated with qualified providers, never treated by the behavior analyst.
  • Prior interventions, their outcomes and earlier hypotheses in the file are information to verify with current data, not confirmed conclusions.
  • Records are obtained with the needed authorization and limited to what is relevant to the referral question; missing records are requested from their sources.
Examples and a practice question: F.1 Reviewing records at the start of a case Practice this: F.1 Reviewing records at the start of a case
F.2

Cultural variables in assessment

Whether a behavior counts as a problem, and which goals make sense, depends partly on the family's values, routines and expectations. Those are part of the assessment, not a courtesy added later. Ask the family directly instead of assuming from a group they belong to. Assess in the language the person and their caregivers actually use, with a qualified interpreter when needed, and read scores from standardized tools cautiously when the comparison group does not resemble the person being assessed.

Example

An assessor notes that a five-year-old eats with his hands at dinner and wants to target fork use. When asked, the family explains that eating with the hands is how meals are eaten at home. The goal moves to something the family does want: sitting with everyone until the meal ends.

Common confusion

The assessor's own habits are not the yardstick for what is normal. Two traps appear in questions: deciding what a family wants from their nationality, religion or background without asking them, and treating cultural information as something to address after the plan is written.

Key points (4)
  • The family's values, routines and expectations shape which behaviors are concerns and which goals are acceptable, so they are gathered during assessment.
  • Assessment is conducted in the language the client and caregivers use, with qualified interpreters when needed, and norm-referenced results are interpreted cautiously when the norms do not represent the client.
  • The assessor's own cultural norms are not the standard for deciding what is a problem; observations take place in the client's natural routines.
  • Cultural information comes from the family itself, not from assumptions about a group the family belongs to.

Strengths and skill needs

A skills assessment answers two questions: what the person already does well, and which missing skills would make the biggest difference. The second question needs a direct test, not just a report.

Figure 1

Skill deficit or performance deficit?

Does the person perform the skill when tested directly?

Clear instruction, a known strong reinforcer, no competing demands.

No

Possible skill deficit

Check prerequisites and test conditions; gather more direct evidence, then teach any demonstrated skill gaps.

Yes

Performance deficit

Change the conditions: motivation, cues and consequences where the skill should happen.

Successful independent performance shows that the skill occurs under the tested conditions. Nonperformance calls for further assessment of prerequisites, instructions, accessibility, motivation and competing conditions before concluding that the skill is absent.
F.3

Assessing skill strengths and needs

A criterion-referenced assessment compares performance with a list of defined skills, so it tells you which skills to teach next. A norm-referenced assessment compares the person with a reference group of the same age, so it tells you how far performance sits from typical; pick the one that answers the question being asked. Record strengths as well as needs, and check the prerequisite skills a target depends on. A direct test can also separate a skill deficit, where the person cannot do it, from a performance deficit, where the person can but does not under current conditions.

Example

A teacher reports that a teenager cannot tie his shoes. During a direct probe with clear instructions and a preferred consequence, he ties both shoes independently. This demonstrates the skill under those conditions. Next, compare the cues, materials, competing demands and consequences at school before deciding what explains the difference.

Common confusion

An informant saying a skill is present or absent is not the same as measuring it. Also keep the two deficits apart: teaching a skill the person already has wastes time, and adding rewards for a skill the person cannot yet perform does not build it.

Key points (4)
  • Criterion-referenced assessments compare performance with a defined set of skills; norm-referenced assessments compare it with a reference group. The choice depends on the question asked.
  • A skills assessment documents strengths as well as needs and checks the prerequisite skills a target depends on.
  • Direct testing distinguishes a skill deficit (cannot perform) from a performance deficit (can perform but does not under current conditions); informant reports are checked against direct measures.
  • An assessment is evaluated for its fit with the client's age, goals and setting, and is repeated over time to measure progress.

Preferences and reinforcers

Teaching depends on consequences that actually work for this person. A preference assessment finds the candidates; only a test of contingent delivery shows which of them strengthen behavior.

Figure 2

Five ways to ask “which one?”

  • SS Single stimulus

    One item per trial; approach or engagement is recorded.

  • PS Paired stimulus

    Two items per trial, until every pair has been offered.

  • MSWO Multiple stimulus without replacement

    Each chosen item stays out for the rest of the assessment sequence.

  • MSW Multiple stimulus with replacement

    The chosen item goes back into the array.

  • FO Free operant

    Everything stays available; time with each item is recorded.

    Share of time with each item

Chosen or approached Not on the table this trial
Each shape represents one item; rows are successive trials. Positions stay fixed here only to help track items. In an assessment, vary item positions systematically to reduce position bias. These sketches illustrate presentation formats, not the number of trials required. The free-operant card shows time spent with each item while all items remain available.
F.4

Preference and reinforcer assessments

Direct preference assessments differ in how items are presented. Single-stimulus shows one item at a time and records approach or engagement. Paired-stimulus offers two at a time until every pair has appeared. Multiple-stimulus without replacement excludes each chosen item for the rest of that assessment sequence, producing a ranking; with replacement, the chosen item is returned to the array on the next trial. Free-operant observation leaves everything available and records time spent with each item. Choose a format from the person's skills and practical limits, then confirm with a reinforcer assessment that contingent delivery increases responding.

Example

A child who cannot yet choose between objects is shown one toy at a time, and the time he holds each is recorded. Another child who protests when a toy is taken away is observed in a room where all the toys stay available, so nothing has to be removed during the assessment.

Common confusion

A highly preferred item is a candidate reinforcer, not a proven one. Choosing it first does not show that delivering it after a response will make that response more frequent. Preferences also drift, so a ranking from last month may be stale, and a parent's list of favorites is a starting point for testing rather than a substitute for it.

Key points (4)
  • Direct preference assessment formats differ in how items are presented: single-stimulus, paired-stimulus, multiple-stimulus without replacement, multiple-stimulus with replacement, and free-operant observation.
  • The format is chosen by the client's skills and behavior and by practical limits: ability to choose among items, scanning, problem behavior when items are removed, and time available.
  • A preferred item is a candidate reinforcer; a reinforcer assessment tests whether contingent delivery actually increases responding.
  • Preferences change, so they are reassessed regularly; caregiver nominations help choose items to test but do not replace direct assessment.

Finding the function of behavior

Functional assessment asks why a behavior keeps happening: which consequence it produces. Three families of methods answer that question with increasing confidence, and only one of them changes variables to show cause.

Figure 3

Three ways to look for function

Indirect

Data come from
What caregivers and staff report
Anything arranged?
No
Can show
Situations to look at; a first guess about function
Examples
Interviews, rating scales, questionnaires

Descriptive

Data come from
Direct observation in everyday settings
Anything arranged?
No
Can show
Events that go together with the behavior
Examples
Narrative ABC, structured ABC, scatterplot

Functional analysis

Data come from
Direct observation in planned conditions
Anything arranged?
Yes: test and control conditions; synthesized tests combine suspected contingencies
Can show
A functional relation, when replicated test-control differences support it
Examples
Standard, brief, trial-based, latency-based, synthesized

Stronger evidence about function, more time and planning →

Moving right, the methods give stronger evidence about function and usually ask for more time, training and risk planning. Only a functional analysis arranges conditions, so only it can show that a consequence maintains the behavior. The first two generate hypotheses.

Figure 4

The conditions of a standard functional analysis

  • Attention

    Arranged before: The adult is present but busy and not interacting.

    After problem behavior: Brief attention, such as a concerned comment.

    Tests: Social positive reinforcement: attention

  • Escape

    Arranged before: Tasks are presented and prompted continuously.

    After problem behavior: The task stops for a short break.

    Tests: Social negative reinforcement: escape

  • Tangible

    Arranged before: A preferred item is taken away after brief access.

    After problem behavior: The item is given back for a short time.

    Tests: Social positive reinforcement: items or activities

  • Alone

    Arranged before: No one present and nothing to do.

    After problem behavior: No programmed consequence.

    Tests: Automatic reinforcement

  • Play Control

    Arranged before: Preferred items available, frequent attention not tied to behavior, no tasks.

    After problem behavior: No programmed consequence.

    Tests: Comparison for every test condition

The social test rows arrange a suspected establishing condition and a response-contingent consequence. Alone provides no programmed social consequences. Play/control provides preferred activities and response-independent attention without demands; it reduces motivation for the tested social reinforcers rather than eliminating all reinforcement. Compare replicated patterns across test and control conditions. Arrangements vary, and ignore may replace alone when needed for safety or feasibility.

Figure 5

Reading a functional analysis against the control

  • Attention
  • Escape
  • Tangible
  • Alone
  • Play (control)

A. Elevated in one test condition

↑ Responses per minute

Sessions →

B. High everywhere, alone included

↑ Responses per minute

Sessions →

Illustrative data, not a real case. Conditions alternate across sessions. In A, responding is elevated only in the attention condition relative to play, which supports an attention function. In B, responding stays high in every condition, alone included, which suggests automatic reinforcement but calls for further analysis before concluding.
F.5

Descriptive and indirect assessment

Indirect assessment gathers information from people who know the person, through interviews, rating scales and questionnaires; it is fast but only as accurate as the informant. Descriptive assessment means watching the behavior where it happens and recording what comes before and after it, without changing anything: narrative ABC notes, structured ABC checklists or a scatterplot. A scatterplot shows when behavior clusters across the day, which tells you when to look; it does not tell you why. Both kinds of data show what goes together, which is useful for forming a hypothesis but not for proving one.

Example

Over two weeks of structured ABC recording, nearly every outburst in a preschool class is followed by a staff member kneeling to talk with the child. That makes attention a strong guess. But staff also kneel to talk with that child often when there has been no outburst, so the pattern alone cannot show that attention maintains the behavior.

Common confusion

Do not mix up the two names: indirect means someone reports, descriptive means someone observes directly. Neither manipulates anything, so neither is a functional analysis. A consequence that often follows a behavior may follow it just as often when the behavior does not occur, which is why comparing those two probabilities helps.

Key points (5)
  • Descriptive assessment is direct observation of behavior and surrounding events in natural conditions, without manipulating variables (narrative ABC, structured ABC, scatterplot).
  • A scatterplot shows when behavior tends to occur across time periods; it points to times worth examining, not to a function.
  • Descriptive data show correlations, not causes; comparing how often an event follows behavior with how often it occurs otherwise helps form hypotheses to test.
  • Indirect assessment (interviews, rating scales, questionnaires) relies on informant report; it is quick but depends on the informant's accuracy.
  • A useful descriptive assessment needs an operational definition, trained observers and observations scheduled when and where the behavior is likely.
F.6

Functional analysis

A functional analysis systematically compares arranged test and control conditions to evaluate maintaining variables. In a standard social test, restricted attention, demands or restricted item access precedes behavior, and the suspected reinforcer follows it. Alone or ignore conditions withhold programmed social consequences to examine persistence without social reinforcement; the behavior can still produce its own consequences. Play/control typically provides preferred activities, frequent response-independent attention and no demands to reduce motivation for the social reinforcers being tested. It does not guarantee the absence of automatic reinforcement. Brief, trial-based and latency-based formats change how tests are organized or measured; synthesized tests combine suspected contingencies and do not isolate each component. Every format requires appropriate expertise, consent, risk planning and stopping rules.

Example

Across fifteen short sessions, a girl's hand biting stays near zero in play, escape, tangible and alone conditions, but climbs in every attention session, where an adult says "please don't do that" after each bite. Responding is elevated only where attention follows the behavior, which supports an attention function.

Common confusion

Never read a test condition alone: high responding means something only relative to the control. Behavior that stays high in the alone condition, and across the others, points toward automatic reinforcement; elevation in two test conditions suggests more than one function. When nothing is differentiated, the result is inconclusive, not proof of an automatic function.

Key points (5)
  • A functional analysis arranges test conditions that each test one possible function: attention, escape and tangible deliver that consequence after problem behavior, and alone or ignore delivers no social consequence; they are compared with a control (play) condition.
  • Results are read against the control: elevated responding in one test condition indicates that function; elevation in several indicates multiple control; persistent responding in the alone condition and across conditions suggests automatic reinforcement; no differentiation is inconclusive.
  • Variations adapt the method: brief FA, trial-based FA embedded in routines, latency-based FA that ends sessions at the first response, and precursor analyses.
  • A synthesized contingency analysis informed by an interview (IISCA) combines several reinforcers in one test condition against a matched control; researchers debate its efficiency and what it can and cannot isolate.
  • Before an FA, risks are assessed and managed: informed consent, termination criteria, medical input and protective measures as needed.

From data to decisions

Assessment ends in decisions: whether behavior-analytic services are needed at all, what else the person needs, which goals come first and which procedures fit the function that was found.

F.7

Deciding on services and referral

The assessment data decide whether behavior-analytic services are warranted. When behavior or skill needs are socially significant, services are recommended. When the data point to a need outside behavior analysis, such as a medical, hearing, swallowing, motor or mental-health concern, the person is referred to a qualified provider, with consent and the relevant information. A referral can run alongside behavior-analytic services that are still indicated. When no significant need appears, or goals have been met and maintained, services are not started or a transition out is planned.

Example

An assessment for food refusal shows that a toddler coughs and gags on several textures. The analyst refers the family for a swallowing evaluation before any feeding procedure begins, and keeps working on mealtime routines that do not depend on that result.

Common confusion

A referral is not a handoff that ends your involvement, and it does not replace services the data still support. The opposite mistake is keeping a non-behavioral problem inside a behavior plan because it shows up as behavior: if pain might explain the change, the first step is a medical check.

Key points (4)
  • Assessment data show whether socially significant behavior problems or skill needs call for behavior-analytic services.
  • Data suggesting needs outside behavior analysis (medical, sensory, psychological, swallowing, motor) lead to a referral to qualified providers.
  • When data do not show a significant need, or goals have been met and maintained, services are not started or a transition out is planned.
  • A referral is coordinated with consent and relevant data, and does not replace behavior-analytic services that are still indicated.
F.8

Prioritizing goals and procedures from assessment data

Not every possible target is equally important. Give priority to goals that improve safety, open access to new settings and reinforcers, act as behavioral cusps or pivotal skills, and help the person in the long run. Shape goals with the person's own preferences and assent and with the family's priorities. Then match procedures to the function found: a replacement behavior should produce the same reinforcer as the problem behavior, at least as quickly and with no more effort, and be acceptable where the person spends time.

Example

Assessment shows a boy hits classmates to get the ball at recess. Teaching him to say "my turn" earns him the ball, which matches the function, and peers respond to it. A plan that only removed recess would leave the function untouched and take away a setting he needs.

Common confusion

A procedure chosen by the form of the behavior, rather than its function, often fails: time-out for behavior maintained by escape can hand the person exactly what the behavior was getting. Goals picked because they are easy to teach, easy to see or preferred by staff are another trap.

Key points (4)
  • Goals are prioritized by social significance: safety, access to reinforcers and settings, behavioral cusps and pivotal skills, and long-term benefit.
  • Function-based procedures use assessment findings to change antecedents and reinforcement contingencies relevant to the identified function, rather than choosing a plan from topography alone.
  • Goals are client-informed and culturally responsive: the client's own preferences and assent and the family's priorities shape them.
  • A replacement behavior produces the same reinforcer as the problem behavior, at least as efficiently, and is acceptable in the client's settings.

Keep going

Practice the whole chapter in the course

Original questions on every task of area F, with explanations and a record of which tasks you have mastered.

Open the course

Try a sample free, without an account