Direct Observation of Behavior is the Primary Means of Collecting Information about behavior. 

All Assessors are required to collect all data necessary to develop high-quality hypotheses of why the behavior occurs.  The quality of the hypotheses are judged on their ability to make predictions and to find workable solutions.

Students of Functional Behavior Assessment: You are required to obtain information from each of the following sources:

  1. Direct Observation of Behaviors in Context, using Systematic Data Collection Methods and standardized Data sheets
  2. Information from Interviews with Primary Stakeholders using structured or semi-structured clinical interview formats
  3. Review Records of Testing and Prior Intervention
  4. Analyze any permanent products created by the behavior


  1. Direct Observation of Behaviors in Context, using Systematic Data Collection Methods and standardized Data sheets
  2. Information from Interviews with Primary Stakeholders using structured or semi-structured clinical interview formats
  3. Review Records of Testing and Prior Intervention – if available
  4. Analyze any permanent products created by the behavior – if available

Systematic Measurement and Data Collection is required in Functional Assessment.

“Systematic, Direct Observation” is the most reliable way of collecting data on behavior.  Systematic, Direct Observation involves recording incidents or events of behavior as they occur (in life; “in-vivo;” live).

The Functional Analyst observes behaviors in operation on the environment.  The Analyst observes what happens before, during and after the behavior, looking for a “functionally related” patterns of antecedents (events occurring before behavior that have something to do with the behavior), behaviors, and the consequences the behaviors produce in the environment.

  1. You must collect data from a variety of sources
    It is necessary for you to obtain quantitative information regarding initial or “baseline” levels of behavior and to able to collect data throughout treatment.

    • Primary Data must be taken through direct observation of the behavior in the environment.  You must arrange times to observe the environment where the behavior is a concern, so that you, or someone else that you train and equip, can witness and record events of behavior while they naturally occur.
      • When observing the environment, it is best to have a special data recording sheet that organizes and specifies what you will record and how you will record events of behavior.
      • The Data Sheet should give you a very specific (operational) definition of the behavior, so observers can agree upon if and when events occurred or not.
    • Secondary Data must also be collected.  Secondary data consists of information obtained from:
      • Interviews of Stakeholders (e.g., Parents; Teachers; Caregivers; etc.)
        • Formal Questionnaires, Inventories, Clinical Interview forms (Contractors: contact the office)
        • Semi-structured, direct interface or telephonic interview with specific referral concerns addressed systematically
      • Review of Records/Comparison of Data from Permanent Products, Tests, and Reports of Progress

However, before you start your observations, you must clarify the referral concerns. You cannot prepare yourself adequately for doing observations and collecting data systematically until you clarify what it is indeed that you are examining.

Therefore, you start by doing some preliminary interviews with Primary Stakeholders (i.e., anyone that deals with the behavior and is concerned about changing it).

Interview Methods for Determining Referral Concerns

It is important to help those involved understand that the goal is to develop a functional and reasonable set of hypotheses (there is no expectation of complete certainty about why the behavior occurs – we can only come up with the most reasonable hypotheses possible).  This is why you are going to collect data and observe the behavior in contexts in which it occurs.

You help them understand that at this point, you don’t know much about behaviors.

It is good to frame the initial contacts by mentioning that at first, instead of answers, you’re looking for questions.

The best way to start an assessment of Functional Assessment is to generate, along with the Primary Stakeholders, a set of questions that the analysis process sets about to examine.  Here are a few examples:

What are the reasons why Suzanne doesn’t do her homework when she is required to do it?
What would be the best ways in which to encourage Suzanne to cooperate with homework requirements?
Why isn’t Lamar speaking yet?  He is already 5 years old and does not have any hearing impairment.
Is Lamar’s lack of speaking part of a broader deficit in relating and communicating, or is it merely due to language differences or delays?

Based on the information we can gather about the above, what interventions are implied for Lamar’s communication?

Shana engages her parents and teachers in power struggles over simple directions and demands where others her age and others in her classroom typically respond cooperatively?

What are the reasons for Shana’s failures to cooperate and her tendencies to insist on inordinate amounts of control?
What contributes to Shana’s inflexibility?
What are the best ways to help Shana to be more flexible and cooperative?

Tomas is 4 years old and still has no interest in toilet independence.  His parents started trying to teach him more than a year ago.

What are the obstacles involved in teaching toilet independence to Tomas?
Considering the obstacles, what would be the best means of teaching Tomas toilet independence?

Minimizing the Inherent Intrusiveness of Intervention and Data Collection Procedures

Observation and measurement on behavior affects the environment and the behavior to some degree.  While careful not to miss any important information, the Analyst also wants to minimize the effects of measurement on behavior.  For instance, a behavioral recording method that requires frequent entries on a chart can affect the potential for engaging a person in an ongoing, emotional and reciprocal chain of social interaction.

Video is becoming increasingly common and easier to use for behavioral observation.  Video recording of behaviors in the environment can be an accepted form of systematic direct observation – as long as the video is unadulterated.


The following interview tips will help you save time and adhere to the principle of “parsimony,” which is a key tenet of Applied Behavior Analysis or any scientific pursuit for that matter.

Basically, we are looking for the most logical and simplest explanation of the behavior possible.  While we don’t want to overlook anything, we also don’t want to engage in unnecessary assessment time, effort and expense.  We are looking for the most concise, least complex/complicated method of obtaining information, as well as the most parsimonious explanations of the behavior.

Ockham’s Razor

This is an old principle (associated with) that is often misstated as, “the simplest explanation is most likely the correct one.”  More correctly, it refers to the admonition attributed to 14th Century philosopher William of Ockham that, ‘…entities should not be multiplied unnecessarily.’  In science, this is generally applied when two or more competing theories make the same predictions.  The admonition is that ‘…when there are competing hypotheses that adequately explain the data, the simpler one is the better.’

You do not want to waste your time or other people’s time asking irrelevant questions and observing insignificant events.  Here are two interview routines that will help you narrow down the significant referral concerns and help you to set up the most meaningful forms of further data collection.

Making Sure You Cover Everything: Two Interview Techniques the Help Elicit Elaboration and Detail

Direct Observation

This involves collection of information in a way that allows reliable measurement of events of the behavior.  After clarifying referral concerns, you now have an idea of what you are looking for and what you will measure.

  1. Observers first have to agree on what the behavior is and what it is not.  They have to define it very specifically, into something called the operational definition of the behavior.  This prevents separate observers from getting wildly different results when they record events at the same time and in the same place.
  2. Observers have to be sure which aspect(s) or “parameter(s)” (see below) of the behavior are under measurement, and which are not.  For instance, recorders may be interested in collecting data on perhaps, how often a behavior occurs, or how long it lasts, or where and when or under what circumstances does the behavior occur? or some other parameter of behavior that appears relevant and necessary to know in order to figure out why behavior occurs.

The results of direct observations must yield quantitative, numeric data.  While commentary is useful and a good supplement to the quantitative data, measurement can only be reported as a count of something.  Even data obtained from ratings (as in measuring of the severity or the impact of a behavior), are reported as numbers. 

  • Observers should be equipped with data sheets or some method of standardizing observation.  Data sheets can afford  the precision and formatting needed to provide consistent types (parameters) of information.  

Students of Behavior Assessment: Schools and clinics involved in the treatment of persons with special needs and behavioral challenges will usually have someone who manages behavioral assessment.  School Psychologists, Behavior Intervention Case Managers (BICMs), Certified Behavior Analysts (BCBAs; BCABAs), Behavior Intervention personnel are available to school districts and schools.

Contractors: Contact the office.

You should plan several observations in the same setting, as well as observations of behavior in other settings.  To save time, it is good to know what behaviors to look for and to set up the observation with the right kinds of recording forms.  The Functional Analyst often uses specialized behavior recording forms, depending on the behaviors of concern.  If the Analyst knows in advance what types of behavior are of concern, he or she can choose among many different types of recording forms.  I discuss the various different parameters of measurement and the associated types of recordings that are best suited to the particular behavior and parameters being measured.

Referral Sheets, Checklists, and Initial Interviews with Guides or Guides can provide enough preliminary information to let the Analyst know what to look for.  But remember, information derived from these sources does not constitute an adequate observation and subsequently cannot lead to a viable Functional Assessment of behavior.

Referral Sheets, Checklists, and Initial Interviews will be effective in gathering the information you need for the Heading Sections and the basic details of schools and services.

Be clear and open with the person you Interview, or the person(s) who opens their home or classroom to you for the purposes of observing.  Share with them that you are there to observe the IPs behavior – not to judge anyone or to come to any conclusion at the moment.

It is true to the spirit and the purpose of behavior measurement to remain objective and to not reach a final conclusion on data from observations until all sources of information are consulted.  Observation across Environments and Context

BehavioralForm”   Parameters of Measurement

  • In Topographical Terms: Before attempting to count or otherwise collect quantitative information about behavior, it is important to define the behavior very carefully.  This means describing the behavior in a way that other observers would likely agree is indeed an occurrence of behavior.
    • The Analyst defines the behavior in a scrupulous manner proscribed by Applied Behavior Analysis.
    • The Analyst develops a clear and detailed operational definition of the behavior that maximizes potential for “inter-observer agreement.”  If multiple observers are expected to measure the behavior accurately – than a scrupulous definition of the behavior is essential.
  • In Terms of Regularity: The Frequency or “Rate” in which behavior occurs;
  • In Terms of Time: The Duration (how long the behavior lasts), or, Latency (how long it takes for a behavior to begin or occur)
  • In Terms of Impact and Severity: The Intensity or Force of the behavior

Topographical measurement also looks at the generalization or pervasiveness of the behavior…

  • In Terms of Where: Does the behavior occur in some or in a variety of settings?  How pervasive is the behavior in terms of where it occurs?  Does information about where behavior does or does not occur inform us about why it occurs?
  • In Terms of with Whom: Does the behavior occur with specific interactive partners, with certain types of people or roles or with most people?  How pervasive is the behavior in terms of who seems to be associated with it?  Is there something about other individuals’ behavior that influences occurrence of the behavior?
  • In Terms of Specific or Ordinary Circumstances: What type of antecedent events or settings seem to trigger behavior?  What do they have in common.  Do behavioral deficits imply specific FERBs?
Functional Behavioral Relationships

Reminder: The ultimate purpose of the Functional Assessment of Behavior is to form working hypotheses of why the behavior occurs

  • Hypotheses of Behavior can be tested and reformulated as new data come in
  • Hypotheses of behavior help identify skills that can replace referral behavior
  • Hypotheses of behavior can help predict what types of teaching will work

Therefore, while we want information about the forms of behavior that are under assessment, we also want to know – to the best ability that we can, the “function,” (i.e. the “use” or “purposes served”) of the behavior.

Applied Behavior Analysis holds certain standards regarding what conclusions you can make given the data and manner of data collection that you have.  Most of the time, the Analyst performs functional assessment of the behavior – where the outcome of the process should be working hypothesis(es) for the behavior(s) of concern.  ABA maintains that further steps are necessary to form stronger conclusions, such as systematic testing and retesting of hypotheses – which is easier to do in clinical environments, and difficult if not impossible to do in natural environments.

ABA discourages the use of thoughts, intentions or mental states as explanations as to why behavior occurs.

Instead, they look for “contingent” relationships between events that can be observed overtly.  This can be done without necessarily putting forth any judgment as to what the IP thinks or is thinking when behavior occurs,

but the contingency models typically used in the Functional Assessment of Behavior (the “Three- and Four-term Contingencies described below), do indeed imply that consequences become antecedents when the IP “expects” certain consequences to occur when he or she engages in the behavior.

Indeed – contingent relationships are ways of explaining how the IP learns what to expect from his or her behavior.  We maintain that we need to know more about the actual consequences – not only in terms of environmental effects, but also in terms of changes in the IPs perception and other factors known to exert important influence on behavior.

In other words, what one expects is usually a matter of the perception of events the person has – the the way they “connect the dots.”

Perception also governs ways in which the IP finds or has the ability to find workable alternatives to problem behavior – or even whether or not the IP considers the behavior to be a problem.

It is unquestionable that behaviors are inseparable from the effects they have on the environment.  Note that “the environment” includes not only the individual’s surroundings and the people in them, but also the individual under analysis. The individual may behave in ways that produce sensory effects that are not directed or “functionally related” to anyone else or any item, action or task.

Some behaviors don’t have effects in the immediate environment.  For instance, getting instruction, studying and taking tests may occur in different environments and include entirely different chains of actions and effects.  Observation across related environments might be necessary to obtain an adequate analysis of behaviors.

And since the consequences in the environment eventually control behavior (there’s more about how this works in the Functional Assessment sections of the Manual), it is important to observe how different environments “shape” or influence the individual’s behavior by the consequences the behavior produces in each environment.

Finally, important clues can reveal themselves in how the behavior changes from one environment to the next, and importantly, where or when the behavior does not occur.

The Three-Term Contingency

The three-term-contingency looks at the functional relationships or interdependence between antecedents, behaviors and their consequences (ABC).  Theoretically, behavior produces consequences that are either reinforcing (the after, or “consequent” effects make recurrence of the behavior more likely), punishing (the after, or “consequent” effects make recurrence of the behavior less likely), or neutral (they don’t seem to affect the behavior at all).

When a person experiences consequences of a behavior, learning can occur.  When learning occurs, the individual expects certain consequences of the behavior.  In this way, previous consequences are now antecedent conditions – they are part of the knowledge or learning, or expectations or anticipations that make behavior more or less likely to occur.  As we will point out in other articles, this simple linear relationship is far from reality, but complex reality can be organized into a “set” of antecedent and consequent variables that interact in complex ways.

The Four-Term Contingency

The newer term is the “Four-Term Contingency.”  This is much more in line with our point of view. The fourth term, which has to do with all of those other factors that make behavior more or less likely, but that add a great deal of complexity into the mix, has to do with those factors within the individual, such as their history of learning, their developmental capacities, the way their mind and emotions work, their temperament, as well as environmental conditions that also alter how an antecedent or consequence will work from one context to another.

Supplementary Information Gathering and Measurement of Behavior

The Analyst must collect comparison data from other sources before forming any hypothesis of behavior.

Direct observation is necessary to obtain baseline measurement and benchmarks for tracking the progression of behavior.  In this method of Functional Assessment however, data from direct observation is not sufficient alone to form hypotheses about behavior.

The Analyst can collect copious data on the forms or shapes of behavior and its patterns of occurrence.

Very importantly – the Analyst looks for data that show functional relationships existing reliably between the behavior and the consequences the behavior produces.  As pointed out earlier, contingent relationships helps form assumptions or “hypotheses” regarding what consequences the IP “has learned to expect” from his or her behavior.

We go a bit further in examining the most probable perceptions the IP infers from the contingencies of the events – “why the consequences occurred” according to the IP.  We observe and collect data on how other elements, such as arousal and alertness, information, attention and memory processing, etc. affect thresholds for behavior.  For instance, the IP may not respond well to being asked to make a transition.  That may produce consequences of postponing or stopping the transition or the getting the demand to transition changed or dropped.  That contingency is likely obvious to all parties involved – including the IP.

But the problems the IP has with the transition are not apparent.  Information about that may not be readily observable in the environment.  We don’t know what meaning the cue for the transition has or had on the IP.  We don’t know if the IP understood or could benefit from the cues given. We don’t know much about the developmental, skill-based abilities the IP brings to bear in understanding the demand.

Because we observed behavior in action, it doesn’t mean that there is an entire universe of things we still don’t know

There is a universe of other consequences, strategies, possible replacement behaviors, tactics, responses that could work that we don’t know about – and that we really can’t know about simply by observing it.

We don’t know what abilities the person has to figure out what else to do either.   Perception also governs ways in which the IP finds or has the ability to find workable alternatives to problem behavior – or even whether or not the IP considers the behavior to be a problem.

Therefore, direct observation and collection of quantitative or numeric data is important, but it must be balanced with other data.  Other data is necessary to refine and challenge the data collected from observation.  The following is a list of other valuable sources of information that should be consulted if available


Information obtained from interviews with various stakeholders and subjects of the behavior is critical.  This is because, as we will continue to maintain – people respond not to what happens – they respond to what they think is happening.  Interviewing yields insights into others’ hypotheses regarding the behavior.  Interviewees often differ in how they define and interpret the behavior – and their responses to the behavior are likely to reflect if not outright demonstrate (as we predict in this model), their attributions of the behavior.

Given the critical role of interview information, it is important that we do what we can to insure the comfort and openness of the interview.  Active listening is a technique used in the gathering of information that helps insure the comfort-setting and open dialogues needed, as well as a means of insuring the reliability of interview information.  Information obtained from someone who suspects the motives of the Assessor; that feels defensive or judged for some reason; or that is placed in a passive position by the Analyst when the Analyst forms hypotheses prematurely, absent full collaborative discourse with Stakeholder/Interviewees.

Active Listening involves frequent checking with the subject of the interview – to insure that the Analyst has heard or understood the subject correctly.  Active listening also involves addressing what is not spoken, but that needs to be spoken about.

Therefore, when interviewing, you want to be prepared for the interview, but you don’t want to influence the subject of the interview with your own (at this time) premature theories.  You are there to find out their theories.  You are there to get their perspective.  You are not there to persuade them of anything or to ask them if they have tried anything you suggest.  This is a time to listen and clarify and not to make suggestions.

Active Listening

Active listening is really a set of tools the Assessor uses to show to put the interview subject at ease, to get accurate, elaborated and reliable information, and no less important than any other aspect of it – to demonstrate that the Assessor “gets it” when the interview subject expresses something.

When people communicate – they do it on several levels at once.  Active listening looks at layers of context; nonverbal information, and other salient reference points to gain the richest understanding possible.

In this way, the processes of Assessment Consultation are identical to communication methods used in counseling.  Active listening has its roots the fields of counseling and psychotherapy, particularly in Carl Rogers’ “client-centered” therapy.  This is not to say that we are psychotherapists, but rather that the therapeutic technique of active listening, which just happens to come from Rogerian client-centered” therapy, can be very useful in Behavior Assessment and throughout the Consultation relationship.

Active listening means responding to the other’s frame of reference.  Be sure to make a real distinction between merely hearing the words and really listening for the message. Listen for what the person is thinking and/or feeling from the other person’s own perspective. To listen effectively, we must be actively involved in the communication process, and not just listening passively.

The Assessor addresses what the interview subject not only says, but what they mean.  Figuring out what someone else means is an ongoing process that is necessary – while at the same time fraught with potential error.   You probe respectfully, carefully, and thoroughly enough without being intrusive.  You periodically test your understanding of the other by restating or clarifying what the other expresses in your own words, or by showing them your understanding by some other means.  This usually encourages the other to build on the thoughts and feelings he or she has just expressed and to explore further.

As mentioned, there will be error.  You will get it wrong here and there.  By encouraging openness in your words and [listening-focused] actions, and frequently checking in and clarifying for accuracy, you reduce error.

How to Use Active Listening
  • Don’t respond to just the meaning of the words, look for the feelings or intent beyond the words.   The dictionary or surface meaning of the words or code used by the sender is not the message.
  • Inhibit your impulse to immediately answer questions. The code may be in the form of a question. Sometimes people ask questions when they really want to express themselves.  They are not really open to hearing an answer at this time, nor is it the time to try to give answers.
  • If you are confused and know you do not understand, either tell the person you don’t understand and ask him/her to say it another way, or use your best guess. If you are incorrect, the person will realize it and will likely attempt to correct your misunderstanding.
  • Active listening is a very effective first response when the other person is angry, hurt or expressing difficult feelings toward you, especially in relationships that are important to you.
  • Use eye contact and listening body language. Avoid looking at your watch or at other people or activities around the room. Face and lean toward the speaker and nod your head, as it is appropriate. Be careful about crossing your arms and appearing closed or critical.
  • Be empathic and nonjudgmental.  You can be accepting and respectful of the person and their feelings and beliefs without invalidating or giving up your own position, or without agreeing with the accuracy and validity of their views
  • Active listening is of course, the best tip.  Active listening is useful at any and every time in the information gathering process.

There are two other rich sources of data that you should consult

Permanent Products

Permanent products have to do with the lasting effects behaviors produce and are considered another form of reliable evidence.  While examination of permanent products of behavior constitutes a form of direct observation, the analysis usually occurs after the behavior occurs.  Analysis involves the examination of the evidence behavior produces, such as completed worksheets, tests, projects, etc.  It can also include the assessment of property damage, bruises and other evidence of behavior.

IP Records

Examination of records can only be considered supplemental to direct observation.  In fact, since it may bias the potential Observer/Analyst, it is a choice that you make whether or not to look at records before, during, or after the series of direct observations that you will do.

This form of Functional Analysis places particular emphasis on the role of prior learning and in particular, on the ways in which the IP processes information (i.e makes meaning and learns from antecedents and consequences that occur surrounding the behavior).

The very concept of the “Identified Person” or “IP,” implies that someone else identified the problem and made the referral.  This is of course – the Referral Source.  This person usually has or knows of records and histories of evaluation, testing, former behavior plans and intervention reports, etc. that yield valuable insight into the developmental capacities and skill sets the IP possesses.  This information, given that it is current enough and accurate, can shed important light what the individual learns or “takes away” from a set of antecedent and consequent events.