ABA is a science in which we make data-based decisions to demonstrate progress to our

Data are the quantifiable results of our measurement

Frequency of Peter’s Aggression

- IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
- 33 SOMEHOW

Observations and anecdotal reflections can be subjective

- Opinions and anecdotal reflections can be subjective
- Relative perspective
- Inaccurate memories

Data are objective

- Unbiased observation
- Consistent measurement
- In the moment

Step 1: Identify the behavior you are collecting data

- The interaction between a living organism and his or her environment
- Dead man’s test
- Something that an organism does that can be observed and measured through space
and time
- Undesired behavior to be reduced
- Behavior reduction
- Desired behavior to be increased
- Skill acquisition
- How to identify the behavior you are collecting data on
- Behavior reduction – outlined in the behavior intervention plan
- Skill acquisition – outlined in the skill acquisition plan

Step 2: Read and Understand the definition of the behavior you are collecting data on
- Definitions of behavior that tell you what behaviors to observe and exactly which
individual occurrences should be recorded. Good operational definitions are:
- Objective
- Only including directly observable aspects
- Clear
- Anyone can understand
- Complete
- All important info and examples/nonexamples
- Individualized
- Person by person
- We will review more about how behaviors are defined later in our course!
Step 3: Identify the method of data collection being used to track each type of behavior
- Types of Measurement
- Continuous measurement
- Discontinuous measurement
- Permanent product
- We will review more about these types of measurement later in our course!

Step 4: Have necessary materials ready to collect data

- Have necessary materials ready to collect data
- Pen/Pencil and Paper Data Sheets
- iPad
- Clickers and Timers
- Other necessary materials

How are Behaviors Defined?

- Behaviors will be defined based on their topography
- What you can observe with your senses (see, hear, feel)
- Typically include examples and non-examples
- Behaviors will be defined on an individual basis
- It doesn’t matter what the behavior is called
- Make no assumptions about the definition

Data Qualifiers
- May include both acts and attempts
- Act: successfully engaging in the behavior from start to finish
- Attempt: initiates the behavior but does not complete (stops action on own,
blocked by staff, interrupted by an another action)
- May include a more specific qualifier
- How long
- How loud
- How far away

Examples of Operational Definitions

- Self-Injurious Behavior (SIB): Any occurrence of the child hitting head to wall for more
than 1 second when engaging in preferred activities or no demands are being placed.
- Non-example: Any occurrence of head hitting during work/while demands are
being placed.
- Property Destruction (PD): Any act of or attempt to use materials other than how they
were intended (e.g., dumping, throwing, crumbling, tipping over, tearing items)
- Non-example: Any occurrence of similar behaviors with materials as intended
(e.g., throwing a ball while playing baseball)
- Aggression (Agg): Any occurrence of hitting, kicking, and/or pulling the hair of another
person
 - Non-example: Do not count attempts in which the child does not make contact
- Talking out: Any vocalizations during group instruction that are not initiated by the
teacher (e.g., teacher calls on student)
- Non-example: Child talking to other students about science during group activity

Why are Operational Definitions Important?

- Take home point: Collect data only on behaviors in the definition despite occurrence of
other undesirable behaviors.
- More information is always better
- Tell your BCBA
- Write a note in the comments section on the Session Sheet (i.e. data sheet)

Continuous Measurement
Measurement conducted so that all instances of the behavior of interest are observed and
recorded during the observation period
- Pros
- Most precise measure
- Cons
- Most time intensive
- May not be accurate when occurring at high rates

Count or Frequency
- The number of responses emitted an observation (tally)
- Ex: he cried three times

Rate
- A ratio of count or frequency per observation period
- Ex: he screamed 3 times in one hour

Duration
- Length of time between the start of the behavior and the end of the behavior
- Ex: he cried for 32 minutes

Response Latency
- Elapsed time between the onset of a stimulus and the initiation of a response
- Ex: he started crying 30 seconds after mom left
- Stimulus: mom left
- Response: crying
Discontinuous Measurement of Behavior
- Measurement conducted in a manner so that some instances of the behavior may not be
observed and recorded during the observation period. These forms of measurement
commonly involve the use of intervals (a predetermined period of time).
- Pros
- Easier, less time intensive
- Cons
- Less precise
- May overestimate or underestimate occurrences
- Whole-Interval Recording
- Definition: Measuring whether or not a behavior is occurring continuously for the
duration of a predetermined amount of time. (+/-)
- Intermittent or single occurrence of behavior does not count
- Whole interval recording tends to underestimate the occurrence of
behavior
- Intervals of time should be short
- Example:
- He cried for 30 minutes during a 30 minute class. (+)
- He cried for 10 minutes during a 30 minute class. (-)
- Convert data to a percentage of intervals
- There were four 30 minute classes today. He cried for the duration of 1 of
4 classes, or 25%.
- When to use Whole-Interval recording:
- Trying to increase duration of behaviors (e.g., time on task)
- Behaviors occurring continuously for a significant amount of time
- Partial-Interval Recording
- Definition: Measuring whether or not a behavior occurs at least once during a
predetermined amount of time. (+/-)
- Precise number of occurrences is not relevant or measured
- Partial interval recording tends to overestimate the occurrence of
behavior
- Interval of time should be short
- Example:
- He cried once during 30 minute class. (+)
- He cried 372 times during a 30 minute class. (+)
- Convert data to a percentage of intervals
- There were four 30 minute classes today. He cried at least once during 2
of 4 classes, or 50%.
- When to use Partial-Interval recording:
- Not concerned with duration of behavior
- When there are concerns about the accuracy of frequency data that
cannot easily be mitigated through training or shifting of other
responsibilities
 - Momentary Time Sampling
- Definition: Measuring whether or not a behavior occurs at an exact moment of
time (+/-)
- Repeated sampling of data collection at a regular interval
- Example:
- At 9:00, he was crying. (+)
- At 9:42, he was crying. (no data collection)
- At 10:00, he was not crying (-)
- Convert data to a percentage of intervals
- Data was collected on the hour. He was in school for six hours. He was
crying 2 times, on the hour. He was crying 33% of the time sampled.
- When to use momentary time sampling:
- When the behavior cannot be directly observed for an entire interval

Permanent Product Data Collection

- This measurement looks at the effect of behavior on the environment after the behavior
occurs. You may not directly observe the behavior occur.
- Examples:
- Worksheets
- Pictures of rooms
- Videotaping

Data Sheet and Data Tracking

- What do we do with all of the data you collect?
- We graph the data!
- Why do we graph?

Visual analysis
- Quicker and easier to see trends and make decisions
- Easier for parents, doctors, teachers to read

Parts of a Graph
- Y-Axis
- Runs vertical. Indicates a measurement of behavior (frequency, duration, count,
percentage)
- Condition Change/Phase Change Line
- Separates baseline data from intervention data, or one intervention from another
intervention.
- Data Point
- One measure of behavior at one point in time
- Condition/Phase Label
 - Identifies what type of intervention is in place.
- Data Path
- This is the series of data points
- X-Axis
- Runs horizontal. Indicates a measure of time (Calendar date, time, session,
interval)
- Axis Label
- Indicates what you are measuring on the Y-axis (e.g., total frequency per day)
and the measurement of time (e.g., sessions, dates) on the X-axis.
-
Graphing Lingo
- Variability
- The difference between multiple data points in a data set
- Rates of bx can have high/low/no variability (no variability = stable)
- Not a good thing
- Level
- The mean location of a portion of a data set relative to the Y-axis
- Typically described as high, medium, or low
- Trend
- Overall direction of a data path over time.
- Rates of behavior can be increasing, decreasing, or maintaining (stable).

Line Graphs
- This type of graph is used to track changes over short and long periods of time. Line
graphs are frequently used to graph skill acquisition data in the form of percent correct or
percent independent. Line graphs are also used to track behavior reduction data in the
form of frequency, rate, duration, and whole or partial interval data once converted to
percent of occurrence.
- Look at the line graph example, based on the information given would you say the
intervention was successful at reducing the target behavior?
Cumulative Record
- Measure a running total across time.
- Based on the example, as of 5/18, this individual had listened to 0 minutes of a podcast.
As of 5/20, they had listened to 20 minutes TOTAL. That means, between 5/18-5/20,
they listened to 20 minutes. If we jump to 5/23, we see they listened to 70 minutes total.
That means, from 5/18-5/23, they listened to 70 minutes across all of those days.
Bar Graphs
- A diagram that uses the height or length of lines or rectangles of equal width to
represent numerical values of given variables.
- We can use bar graphs to graph data from preference assessments as well as functional
behavioral assessment tools (e.g., QABF).