Part C- Measurement, Data Display, and Interpretation
C-1: Establish Operational Definitions of Behaviour
Operational Definitions: A good definition should be objective, clear, and complete. It should describe the beginning and end of an instance of behaviour, as well as describe what is not an instance of the target behaviour.
Function- Based Definition: The function of the behaviour is most important in these definitions. Responses are part of a designated response class, dependent on their typical effects on the environment. These definitions tend to be simpler.
Topography-Based Definition: The shape or form of the behaviour is most important in these definitions. They should only be used when a specific function has not been identified for the target behaviour or when each instance of the target behaviour does not produce the typical outcome or the outcome may be produced by other target behaviours or events.
C-2: Distinguish Among Direct, Indirect, and Product Measures of Behaviour.
Direct Measures: Observing and recording target behaviours and other measurable information. This type of measure is the most accurate and useful.
Indirect Measures: Data obtained through methods other than direct observation, such as interviews and checklists. This type of measure also gather useful information, but it does so through interactions with others. It should be used in combination with direct measures, never as the only measurement of behaviour.
Product Measures: Examining the complete effects that the behaviour had on the environment.
C-3: Measure Occurrence (e.g., count, frequency, rate, percentage)
Count: Number of occurrences of a behaviour.
Example: 7 sips of wine
Frequency/Rate: Number of occurrences of behaviour per unit of time.
Example: 7 sips of wine/30 minutes
Celeration: Think of acceleration or deceleration. Celeration measures how response rates change over time.
Percentage: Ratio of events per total number of events.
Example: A clinician sitting for the BACB exam receives a score of 480/500, therefore their percentage will be 96%. (This is going to be you!!!! Sending lots of good vibes!!!)
C-4: Measure Temporal Dimensions of Behaviour (e.g., duration, latency, inter response time)
Temporal Extent: All instances of a behaviour occurring over some amount of time
Duration: How long the behaviour occurs
Temporal Locus: All instances of a behaviour occurring over some amount, in relation to other events
Response Latency: The time between the presentation of a stimulus and the initiation of the response
Inter-Response Time (IRT): The time between two occurrences of the behaviours within the same response class
Short IRT: High rates of responding
Long IRT: Low rates of responding
C-5: Measure Form and Strength of Behaviour (e.g., topography, magnitude)
Topography: Physical form or shape of the behaviour; it’s measurable and can quantify the behaviour
Example: Hitting head versus playing with a fidget toy
Magnitude: The force or intensity of a behaviour
Example: Speaking versus screaming
C-6: Measure Trials to Criterion
Trials to Criterion: Number of responses to reach mastery (predetermined level of behaviour)
Example: 80% responses over 2 consecutive days
C-7: Design and Implement Sampling Procedures (i.e., interval recording, time sampling)
Interval Recording/Time Sampling: Observation of behaviour occurring or not occurring during a specific time period.
C-8: Evaluate the Validity and Reliability of Measurement Procedures
Validity: Data is directly relevant to the phenomenon measured and reasons for measuring it
Indirect Measurement: The target behaviour is measured in a different way, than designed
Measurement Artifacts: Something that appears to exist because of the measurement or examination methodology
Discontinuous Measurement: Some of responses of the target behaviour are not detected
Poorly Scheduled Measurement Periods: Non-standardized methodology of data collection, where an equal opportunity is not present to observe the target behaviour
Accuracy: The extent to which the observed value matches the true value, or the event as it exists in nature
True Value: A measure obtained by procedures that vary from the procedure in-use and the researcher must take precautions to ensure all sources of error have been considered
Measurement Bias: Error in measurement, likely to be in one direction
Reliability: Consistent measurement
Threats to Reliability
Poorly Designed Measurement System: Unnecessarily difficult or cumbersome one to use
Inadequate Observer Training: Lack of observer training
Observer Drift: Data collection methodology is unintentionally changed
Observer Expectations: The observer’s expectation of how the target behaviour should occur during certain conditions
Naive Observer: The observer collects data, but is unaware of the study’s purpose or experimental conditions
Observer Reactivity: Observer’s awareness that other’s are evaluating the data that is being collected
Calibration: Procedures used to evaluate the accuracy of a measurement system
Inter-observer Agreement (IOA)
Event Recording IOA
Timing-Based IOA
C-9: Select a Measurement System to Obtain Representative Data Given the Dimensions of behaviour and The Logistics of Observing and Recording
C-10: Graph Data to Communicate Relevant Quantitative Relations (e.g., equal-interval graphs, bar graphs, cumulative records)
Equal Interval Graph: These are graphs that display two consecutive points, which represent the same value on both X and Y axis.
Scatterplots: A graph with unconnected points that show relative distribution of individual measures, in relation to the x and y axis. It helps to determine if there is a temporal relation between two individual measures. It may help clinicians determine if there is a correlation between time of day and the behaviour’s occurrence.
Line Graph: Most commonly used graph in ABA; It’s used to communicate a variety of different relations:
Single data path. Example: Frequency of hitting.
Multiple dimensions of behaviour. Example: Prompted vs unprompted requests by an individual
Multiple topographies of behaviour. Example: Chair tipping vs Screaming frequencies
The same behaviour under different conditions. Example: Functional Analysis Data with different conditions
Comparing multiple individual’s data. Example: Frequency of independently greeting adults- John vs Joan
Changing values of an independent variable. Example: Y=decrease in number of meal times and X= binging behaviours during mealtime
Example: Line graph.
Cumulative Records: Illustrate the total number of recorded responses over time and can reveal intricate relations between behaviour and environmental variables; They are typically used when the behaviour occurs only once per observation session.
Example: Cumulative Record graph displaying number of acquired programs per month for an individual.
Semilogarithmic Charts (Standard Celeration Charts): These graphs look at proportional/relative changes in behaviour; used in precision teaching.
Example: Standard Celeration Chart for learning SAFMEDS.
Bar Graph: They are useful in illustrating sets of data that are unrelated, but share a common dimension by which they can be displayed on the x axis. Example: ACT Quantitative Analysis Scale
Example: Bar Graph to display data from weekly AQAS.
C-11: Interpret Graphed Data
Visual Analysis: Interpreting data displayed in a graph. It helps to identify if the behaviour changed in a meaningful way and if that change can be attributed to the independent variable.
Level: The value of data points on the y-axis of a graph. There are 3 levels, high, moderate, or low. How I like to look at the level is by determining the highest data value and seeing if most points are closest to the highest value, at the mid point, or low point.
Variability: This is how much the data moves on the graph. There are 2 types, high variability and low variability.
Trend: The direction of the data path. There are 3 levels, increasing, decreasing, or zero trend.
