Displaying and Interpreting Behavioral Data Flashcards
- Medium with which the behavior analyst works.
- Results of measurement.
- Empirical basis for decision-making.
- It is difficult to assess what is happening with the target you are trying to measure if you only use raw ______; this is why ABA uses graphs.
Data
- Visual format for displaying data.
- Reveals relations between a series of measurements and relevant variables.
- Helps people make sense of quantitative information.
- How behavior analysts organize, store, interpret, and communicate the results of our work.
Graphs
Three (3) Purposes of Graphs: CAID
- Communicate: Communicates our data.
- Assess: Helps us to assess data correctly.
- IV/DV: Shows how the DV and IV are related to each other.
- ) Gives you an immediate picture of an individual’s behavior.
- ) Allows you to explore interesting variations in behavior as they are occurring.
- ) Acts as a judgmental aid to help you interpret the results of a study or treatment.
- ) Acts as a conservative method for determining the significance of behavior change, because a behavior change that is statistically significant may not look impressive on a graph.
- ) Allows for an independent judgement and evaluation of the data.
5 Benefits of Graphs
5 Types of ABA Graphs: LBCSS
- Line Graph
- Bar Graph
- Cumulative Record
- Scatter Plot
- Standard Celeration Chart
AKA: Arithmetic Charts; Add-Subtract Charts
- Graphs in which the distance between any 2 consecutive points on BOTH THE X-AXIS AND Y-AXIS are always the same.
- All intervals are the same size.
Line graphs, scatter plots, bar graphs, and cumulative records are: _________ ____________.
Equal-Interval Graphs
Logarithmic scales, including semi-logarithmic scales, one of which is the standard celeration chart, look at behavior through proportionate or relative change.
They are examples of: _________ _________ ______.
Non-Equal Interval Graphs
AKA: Frequency Polygons
- Most common used graphs in ABA.
- Based on the CARTESIAN PLANE.
- Each point on _____ _______ shows the level of some quantifiable dimension of the DV in relation to the IV in effect when the data was recorded.
Line Graphs
- 2-dimensional area formed by 2 perpendicular lines that intersect.
Cartesian Plane
Comparing data points lets us examine:
LEVEL, TREND, and VARIABILITY
Use _______ _________ when you want your data to effectively communicate the following relevant quantitative relations:
- Data that can be scaled along some dimension, such as time or the order of responses in a sequence.
Line Graphs
Seven (7) Parts of A Line Graph
- Horizontal Axis (X-Axis or Abscissa) (Line Left _ Right)
- Vertical Axis (Y-Axis or Ordinate) (Line Up | Down)
- Condition Change Lines
- Condition Labels
- Data Points
- Data Path
- Figure Caption
AKA: X-Axis or Abscissa (Line Left _ Right)
- Represents passage of time and the presence, absence, or value of the IV.
- Left to right passing of time in equal intervals.
- Tic marks are placed on this with equal spacing between them.
- Utilizing a scale break (= or //) to represent discontinuities in time (times was not taken due to some reason)
Horizontal Axis
AKA: Y-Axis or Ordinate (Line Up | Down)
- Represents the full range of values of the DV (i.e., quantifiable aspect of the target behavior).
- On an equal-interval graph, the scaling of this axis is really important to see changes in the level, trend, and variability in the data.
Vertical Axis
Intersection of the horizontal and vertical axis.
- Usually represents the zero value of the DV.
- Should be marked at zero.
Origin
The vertical lines drawn upward from the x-axis to show points in time at which changes in the IV occurred.
Solid Lines = Major Changes
Dashed Lines = Minor Changes
Condition Change Lines
A label, written at the top and parallel to the x-axis, that describes the experimental conditions in effect during each phase of research.
Condition Labels
Has 2 meanings:
- A quantifiable measure of the target behavior recorded during a given observation period.
- The time and/or experimental conditions under which that particular measurement was conducted.
Data Points
Always displayed as (x, y)
Coordinates
- Connects successive data points with a straight line. Illustrates level and trend of behavior between 2 consecutive data points.
- The ____ ______ should be examined to interpret graphs.
- A maximum of 4 different ______ _______ can displayed effectively on 1 set of axes.
Data Path
- Concise statement that provides information to identify the IV and DV. Also explains symbols used an unplanned events.
- Printed below graph.
Figure Caption
- Line graphs can be vary complicated, especially when multiple data paths are represented.
- Line graphs that are more complex than what we typically may utilize:
1. Two (2) or more DIMENSIONS of the same behavior.
2. Two (2) or more different BEHAVIORS.
3. Measure of the same behavior under DIFFERENT CONDITIONS.
4. Changing VALUES of the IV.
5. Same behavior of two (2) or MORE PARTICIPANTS.
Line Graph Variations
AKA: Histograms
- Based on the CARTESIAN PLANE, similar to the line graph.
- However, there are NO distinct data points representing successive response measures through time.
- Does NOT allow for analysis of variability and trend in behavior.
CANNOT BE USED WITH TIME.
Bar Graph
Use Bar Graphs When You Want Your Data To:
Effectively communicate the following relative quantitative relations:
1. Displaying separate sets of data the ARE NOT related to each another.
Developed by Skinner to record data in EAB research in 1957.
- You keep adding on responses during each observation period to the total number of all previously recorded responses.
- The y-axis value represents the total number of responses recorded since the very start of data collection.
- Exception: When the total # of responses exceeds the upper limit of the y-axis scale (in which case the data path resets to zero on the y-axis and begins its rise again).
Cumulative Records
- Displays cumulative data.
- Enables a subject to draw his/her own graph.
Cumulative Recorder
Generally Used for Rate/Frequency data.
- The last data point on this graph is how much an individual has acquired at the current time.
Cumulative Record
2 Types of Cumulative Record Response Rates:
- Overall Response Rate
2. Local Response Rate
An average rate of response during periods of time smaller than that for which an overall response rate has been given.
- Same calculation as the overall response rate, but only using a small portion of the data on the graph.
Local Response Rate
An average rate of response over a given time period, such as during a specific session or phase in a study.
- Calculated by dividing the total # of responses recorded during the period by the total # of observation periods indicated on the x-axis.
Overall Response Rate
What relevant quantitative relations are effectively communicated on the cumulative record?
(In other words…..)
-Why would a cumulative record be used rather than a noncumulative graph?
- The target behavior can be measured in cumulative units.
- The target behavior only occurs once per observation period.
- The cumulative record shows how rapidly or slowly the target responses are repeated.
- The cumulative record can be used as personal feedback.
- The effects of the IV are easier to detect on a cumulative record rather than a noncumulative graph.
AKA: Ratio Chart; Multiply-Divide Chart
LOGARITHMIC SCALES look at behavior change through PROPORTIONAL or RELATIVE change.
- X-Axis = In equal intervals
- Y-Axis = Scaled LOGARITHMICALLY
Graphs in which 1 axis is scaled proportionally.
All behavior changes of equal proportion are shown by equal vertical distances on the vertical axis.
-Data that is shown as an exponential curve on an equal interval chart is a straight line on a ________ ________.
Semilogarithmic Charts
A type of SEMILOGARITHMIC CHART.
- Created by Ogden Lindsley to be used in an ABA educational methodology called Precision Teaching.
- Academic and social behaviors are charted.
- Provides a standardized mans of charting and analyzing how frequency of behavior changes over time.
- The scale goes by multiples (2s, 10s, 100s, etc.)
- Students self-monitor their progress by recording data that makes a graph that displays the number of items they performed correctly and the number of errors they made within fixed periods of time distributed across the day or week.
- GOAL: To increase the # of correct answers and decrease the number wrong within the set time.
- Allows data to be squeezed into progressively tighter and tighter bundles.
Standard Celeration Chart
What relevant quantitative relations are effectively communicated on the ration chart?
(In other words…….)
-Why Would I Use This?
Primarily when your concern is promoting RATE OF RESPONDING.
-Rate of responding is really important because research shows that the more rapid and fluent the rate of correct responding, the more durable the learning.
Show relative distribution of individual measures in a data set.
- Data points are UNCONNECTED.
- Depict changes in value on 1 axis correlated with changes in value on the other axis.
- One variable (usually the time of day) is plotted on the y-axis and a second variable (usually days) is plotted on the x-axis.
Scatter Plots
Use Scatter Plots when you want your data to effectively communicate the following relevant quantitative relations:
- Temporal distribution (TIME) of the behavior.
2. The grouping of the individual data points may help to identify elusive environmental stimuli.
Make sure you choose the method that demonstrates the most ETHICAL and VALID representation of the target behavior.
-Do NOT choose a method that skews the target behavior.
Base Decision-Making on Data Displayed in Various Formats
Three (3) Fundamental Properties of Behavior Change: LTV
- Level
- Trend
- Variability
Value on the vertical axis around which a series of data measures converge.
- A change in ______ is illustrated when the data’s average value changes.
- An analysis of ________ answers the question, “How much has the behavior changed?”
- _________ in your data are examined by looking at your data’s mean, median and/or range.
Level(s)
Horizontal line drawn through the data points on the vertical axis equaling the AVERAGE or MEAN value of the data.
- Can obscure important variability
- Adds an easy-to-see summary of average performance.
Mean Level Line
Horizontal line drawn through the data points on the vertical axis that shows the most typical performance within a condition.
-Better than the mean level line when your data has extreme outliers, either high or low.
Median Level Line
OVERALL direction taken by a data path
- The general direction and rate of increase or decrease in which data move over time.
- An analysis of ______ answers the question, “In what direction is the change headed?”
Described in terms of their:
- Direction: Increasing, decreasing or zero _____
- Degree: Gradual or steep
- Extent of variability of data points around the _____.
Trend
LINE OF PROGRESS: A straight line drawn through the data to show the trend.
AKA: Split-Middle Line of Progress
Trend Line
- Freehand (not very accurate)
- Mathematical formula called, the Ordinary Least-Squares Linear Regression Equation (time-consuming and complicated)
- Split-Middle Line of Progress (Best)
How Do We Draw Trend Lines?
- Count (how many data points are on the graph)
- Divide (draw a vertical line and divide the number of data points in half; left & right equally)
- Mid-Rate (find middle point up/down; mid value for the y-axis; if there are even #’s, use the mean of the two middle-most points)
- Mid-Date ( For each half of the data, find the middle points, left to right)
- Quarterly-Intersect Line of Progress (Connect the 2 mid-date & mid-rate points of intersection)
- Split Middle line of Progress (Shift the quarterly-intersect line up or down; keep it parallel to itself so that an equal # of data points fall above or below it.
Six (6) Steps To A Split-Middle Line of Progress
- The extent to which data “bounce around” on the graph.
- An analysis of __________ answers the question, “How consistent is the change that is taking place?”
- Frequency and degree to which multiple measures of behavior yield different outcomes.
High degree of _________ = Little to no control over the factors influencing behavior.
Variability
Examining data within each condition, determine the level, trend and/or variability in each condition.
Comparing the data in the different conditions, determine whether change in level, trend, and/or variability occurred and to what extent any changes were significant.
Visual Analysis Of Temporal Relations Of Data Within And Between Conditions