An Overview of Data-Based Decision-Making as Part of Positive

							Positive Approaches to
      Behavior Change



Richard J. Cowan, Ph.D., NCSP
           Kent State University
Overview
   Defining Positive Behavior Supports
   PBS at the Individual Student Level
   Problem Solving Sequence
   Defining Target and Replacement Behaviors
   Data Collection Systems
   Analyzing Data to Make Informed Decisions
PBS Defined
   “PBS is based on a problem-solving model and
    aims to prevent inappropriate behavior through
    teaching and reinforcing appropriate behaviors.”
   “PBS…offers a range of interventions that are
    systematically applied to students based on their
    demonstrated level of need, and addresses the
    role of the environment…[in the] development
    and improvement of problem behaviors.”

            (Kincaid and Algozzine, 2008)
PBS Defined
   PBS involves the assessment and reengineering of
    environments so people with problem behaviors
    experience reductions in their problem behaviors
    and increased social, personal, and professional
    quality in their lives.
   PBS is a collaborative, assessment-based process
    to develop effective, individualized interventions for
    individuals with challenging behaviors. Support
    plans should focus on proactive, educative and
    functional approaches.
                         (Horner, 1999)
PBS at the Individual Level

   How do we conceptualize an individual’s
    strengths and needs within the PBS framework?
   What information do we need?
   What do we do with this information?
   How can we help the student be successful and
    have a POSITIVE LEARNING EXPERIENCE?
PBS at the Individual Level
   Problem-Solving Sequence
     Problem Identification

       Behavior Definition, Data Collection

     Problem Analysis

       Data Analysis, Develop Intervention

     Treatment Implementation

     Treatment Evaluation
Problem Identification
   Define the behavior of concern (i.e., the target
    behavior) in concrete, observable terms.
   In order to teach the child how to be more
    successful, define an alternative/replacement
    behavior in concrete, observable terms.
   When defining the target and replacement
    behaviors, consider…
     Inclusionary and Exclusionary Criteria
     Make definitions clear, concise, and objective
     Dead man’s test (Maag, 1999)
     Stranger test (Maag, 1999)
Problem Identification
   Develop a data collection system that allows
    the team to determine the frequency, intensity,
    and/or duration of the behavior.
   Make it simple, reliable and valid
   Baseline phase data will include:
     Behavior specifics (frequency, intensity,
      and/or duration)
     Observation of antecedents and
      consequences surrounding behavior to assist
      in determining function of behavior (for
      intervention planning)
       Measuring
Behavior Specifics
Types of Measurement
Techniques
   Permanent products
   Direct Observation of Behavior
       Event recording
       Duration recording
       Latency recording
       Interval recording
           Partial interval
           Whole interval
           Momentary time sampling   (Alberto & Troutman, 2006)
Permanent Product
   Use it when
       You can see evidence that the behavior happened
       You are interested in how many times the behavior occurs
   Keep in mind
       Each time the behavior happens, it should result in the same
        end-product
       Only the target behavior should be able to have produced
        the evidence
   Not good to use if
       The behavior does not result in a product
       You are interested in how long the behavior occurs
Using Permanent Product
Measurement
   Determine what constitutes an acceptable
    outcome of the behavior—use the conditions
    established in the behavioral definition
   Evaluate the outcome of the behavior
   Make note of how many of the products were
    produced acceptably according to the
    definition of the behavior
Event Recording
   Use it when
       You want to know how many times a behavior happens
   Keep in mind
       The behavior must have a definite beginning and ending
   Not good to use if
       The behavior occurs too quickly to count each time it
        happens accurately
       You are concerned with how long the behavior occurs
        instead of how many times it happens
Duration Recording
   Use it when
       You want to know the length of time a behaivor occurs
   Keep in mind
       The behavior must have a definite beginning and ending
   Not good to use if
       The behavior occurs too quickly to measure the duration of
        each occurrence
       You are concerned with how many times the behavior
        happens instead of how long it occurs
Latency Recording
   Use it when
       You want to know how much time passes between an
        opportunity to perform a behavior and the initiation of the
        behavior
   Keep in mind
       The behavior must have a definite beginning and ending
   Not good to use if
       You are concerned with how many times the behavior
        happens or the total duration of the behavior instead of
        how long it takes to initiate the behavior
Interval Recording
   Use it when
       You want to estimate how many times a behavior happens
       You can not count each occurrence of behavior; instead
        the number of intervals that the behavior happened within
        a period of time is counted
   Keep in mind
       Good to use if the behavior happens quickly
   Not good to use if
       You want to know the actual number of times a behavior
        happens or how long a behavior occurs
Types of Interval Recording
   Partial interval recording
       Did the behavior happen at all during the interval?
   Whole interval recording
       Did the behavior happen for the entire interval?
   Momentary time sampling
       Was the behavior happening at the end of the
        interval?
                     What do you want
                      to know about
                       the behavior?




Number of Times                         Amount of Time




Permanent Product
                                           Duration
       Event
                                           Latency
Interval Recording
Practice: How would you measure
these behaviors?
   The number of times Mike talks out.
   The amount of time Mariah spends twirling her hair.
   How long it takes Jamaal to put his heel on the floor
    when asked to stop toe-walking.
   The number of times Jack flaps his hands.
   How long it takes Jose to make eye contact after
    being asked to do so.
   The length of time Ariel spends tantrumming.
   The number of times Rachelle puts her hands in her
    mouth.
   The number of blocks Jason stacked in the color
    sequence.
     Measuring
Antecedents and
 Consequences
Conditions Surrounding
Behavior
   Setting Events
     Temporally or contextually removed stimuli that bear a
      functional relationship to the behavior (Wahler & Fox,
      1981)
   Antecedents
     What happens just before the behavior occurs?

     What “sets the stage” for a behavior?

   Consequences
     What happens immediately after the behavior occurs?

     What message does the child receive re: the behavior?

   Look for patterns over time!
Conditions Surrounding
Behavior
   Consequences of Behavior
       Positive Reinforcement (access to something
        desirable)
         object, activity, attention, social engagement,
          sensory stimulation (i.e., it feels good)
       Negative Reinforcement (escape from something
        undesirable)
         difficult task, physical discomfort, social
          embarrassment
       Punishment
Problem Analysis
   Generate hypotheses regarding what might
    be causing or maintaining the target
    behavior? (To decrease disruptive behavior)
   Under what conditions is the student
    successful? (To enhance success)
   Generating such hypotheses…
       Provides a summary of FBA findings
       Offers potential explanations as to what may be
        causing and/or maintaining a behavior
       Guides in the development of an intervention
Developing Positive Behavior
Supports
   Punishment stops behavior!
   The problem with using punishment alone is
    that it fails to teach the student what to do!
   Side-effects associated with punishment:
       learned helplessness,
       emotional reaction,
       increased aggression (especially if modeled), and
       response substitution (e.g., replacing biting with
        hitting)
Developing Positive Behavior
Supports
   So how do we teach behavior?!
   We provide multiple opportunities to practice
    new skills
   We support learning through modeling,
    coaching and prompting (as needed)
   We reinforce desired behavior (so the child
    knows what is expected/desired behavior)
   How do we reinforce good behavior?
Developing Positive Behavior
Supports
   Positive reinforcement (something added that
    increases the likelihood of behavior)
       Positive/negative attention from adults, peers
       Access to tangibles and/or preferred activities/events
       Sensory feedback (i.e., it feels good)
   Negative reinforcement (an aversive stimulus is
    removed, thus increasing likelihood of behavior)
       Escaping/avoiding tasks, demands, social interactions,
        certain settings, etc.
   Differential Reinforcement (the “BOGO!” intervention)
Developing Positive Behavior
Supports
   Differential Reinforcement: What is it?!!
       The combined application of positive reinforcement
        and extinction
       Extinction is the elimination of reinforcement from a
        previously reinforced behavior
       The result is the disappearance of the
        inappropriate behavior and an increase of the
        appropriate behavior
       DRO (Omission), DRA (Alternative), and DRI
        (Incompatible).
Converting, Graphing
  and Analyzing Data
Converting Data
   It is important to ensure that all data collected
    are comparable to one another
       Equal opportunities to display the behavior
           Same number of trials
           Same amount of time observed
       Equalize the data in relation to one another
           Percentage
           Rate
Ways to Report Data
   Number of occurrences, intervals,
    seconds/minutes/hours
   Percentage of occurrences, intervals,
    seconds/minutes/hours
   Rate of occurrences
Practice: How would you measure the
behavior and convert the data?

   How many flash cards a student answers
    correctly
   How long a tantrum lasts
   How many intervals the child was engaged
    with a peer
   How many times the child answers a
    question posed by the teacher
Graphing Data
   Visual representation of current level of
    performance (baseline)
   Visual representation of level of performance
    with intervention in place (treatment,
    intervention)
   Comparison between performance during
    baseline and intervention conditions tells us
    the extent to which the intervention is
    effective
Parts of a Line Graph
   X-axis
   Y-axis
   Labels
   Increment marks
   Data points
   Data path
   Break in Data
   Legend
   Title
Criteria for Analyzing Data
   Number of data points
   Stability of data
   Mean level
   Trend in the data
   Immediacy of change
Number of Data Points
   Are there enough
    data to get a good
    representation of
    the level of
    performance?
Mean Level
   Mean levels report how the behavior occurs
    on average.
   Mean levels tell us what level of performance
    we can expect the behavior to continue at if
    we do not intervene.
Calculating Mean Levels
   Add the value of each
    data point in the phase
   Divide by the number of
    data points in the
    phase
   Indicate mean level on
    graph with a solid
    horizontal line within
    the phase
Stability of Data
   Do the data represent a consistent level of
    performance or are the data variable and
    unpredictable?
   Stable patterns of behavior allow us to predict
    behavioral responses more consistently.
Calculating Stability of Data
   Calculate mean level
   Stable data have
    points that are within
    50% of the mean
   Variable data have
    points that are
    greater than 50%
    away from the mean
Trend of the Data
   Trend suggests whether the behavior is
    generally accelerating, decelerating, or
    staying about the same in strength or
    occurrence.
   Knowing the trend of the data, we can predict
    where the behavior will be over time without
    intervention.
Split-Middle Estimate of Trend:
Step 1
                                         Mary's CBM Reading Performance

                        70
 Words Read Correctly




                        60
                                                                 Divide data in half
                        50
                        40
                        30
                        20
                        10
                        0
                             1   3   5    7   9   11   13   15     17   19   21   23   25   27   29
                                                            Weeks
Split-Middle Estimate of Trend:
Step 2
                                         Mary's CBM Reading Performance

                        70
 Words Read Correctly




                        60
                                                                 Find the middle or median
                        50
                                                                 score for each half
                        40
                        30
                        20
                        10
                        0
                             1   3   5    7   9   11   13   15     17   19   21   23   25   27   29
                                                            Weeks
Split-Middle Estimate of Trend:
Step 3
                                         Mary's CBM Reading Performance

                        70
 Words Read Correctly




                        60
                                                                 Mark the midpoint in time
                        50
                                                                 for each half
                        40
                        30
                        20
                        10
                        0
                             1   3   5    7   9   11   13   15     17   19   21   23   25    27   29
                                                            Weeks
Split-Middle Estimate of Trend:
Step 4

                                         Mary's CBM Reading Performance

                        70
 Words Read Correctly




                        60                                       Connect the two midpoints to
                        50                                       create the trend line
                        40
                        30
                        20
                        10
                        0
                             1   3   5    7   9   11   13   15     17   19   21   23   25   27   29
                                                            Weeks
Immediacy of Change
   How quickly the behavior responds to
    implementation (or withdrawal) of intervention
    gives us an indication of the strength of the
    intervention.
Determining Immediacy
   Analyzing the
    presence or
    absence of an
    immediate
    separation of data
    from that
    observed in the
    previous phase
   Make relative
    comparisons
Comparison of Behavior Across
Conditions
   Comparison between baseline and
    intervention phases
   Extent to which behavior change represents
    a meaningful change
   Determining whether changes in intervention
    are indicated
Data-Based Decision Making
   How is the student performing as compared
    to goals established during Pre-Consultation
    and Problem Identification?
   How is the student performing as compared
    to his/her peers in the classroom setting?
       How do you develop local norms?
       What is “meaningful” for this student?
       What is “reasonable” for this student?
Data-Based Decision Making
   How will you maintain treatment gains?
   How will you generalize the treatment gains?
   How and when will you fade the intervention?
       Is it feasible to maintain the intervention?
   Other considerations?

   Critical Resource: Alberto & Troutman (2007)
    Applied Behavior Analysis for Teachers