Using Data and Research Methods in Behavior Analysis
How we use data:
- An Intervention is a program or period of time in which action is taken to alter an existing
situation, such as a target behavior.
- Baseline has two meanings: a) it can refer to the data collected before the intervention begins
or to the period of time during which those data were collected.
- variables- characteristics of people.
- A Table is a systematic arrangement of data or other information in rows and columns for easy
- A Graph is a drawing that displays within a set of data, typically showing how one variable
changed with changes in another variable.
Types of gaphs:
Line graphs: use straight lines to connect successive data points that represent the
intersects of plotted values for the variables scaled along the horizontal and vertical axes / most
common type / the horizontal axis typically scales chronological time or sessions.
Bar graphs: use vertically arranged rectangles to represent data points scaled along the
Cumulative graphs: or cumulative records are line graphs are line graphs in which the
measure of behavior accumulates across units scaled along the horizontal axis. /the graph is
smoother and the line can never turn downward.
- A complete graph should have 5 points :
1) Axes: vertical and horizontal
2) Axis scaling and labels: *with measurement units
3) Data points:
4) Phase lines and labels: line and cumulative graphs are usually divided into two or more
phases, separated with vertical lines (usually dashed). The phases usually represent treatment and
no-treatment periods, which are correspondingly labeled “intervention” and “Baseline”
5) Caption: A figure caption begins with its figure number, which reflects the order in which the
graph is reffered to in the text material of the report or book.
Using Graphs and Basic Research Methods:
Graphic Analysis: in which behavior analysts inspect graphed data to evaluate whether the
behavior changed substantially when intervention techniques were implemented.
- judging whether the program is working after a week or so of an intervention involves
assessing two trends : a) whether the behavior improved from baseline to intervention. b) has
continued to improve across time during the intervention.
- to help analysis an unclear graph we can draw a trend line: a straight line that best fits or
represents all of the data points in a time period, through or very near the data point for the
- issues with trend lines , first they are positioned correctly, each one looks like it carves the
corresponding data points in half. Second, not appropriate to have a trend line that includes a scale break. Third, more complex procedures are available for constructing and interpreting trend
lines if higher precision is needed.
- The lower the overlap from baseline to intervention, the greater the interventions effect
- the ideal condition has no overlap
- Behavioral excess: the lowest data point in baseline would have a higher value than the highest
data point in intervention.
- Behavioral deficit: the highest data point in baseline would be lower than the lowest data point
- Difficulties in evaluating trends can arise from data problems of three types: excessive
variability, a decreasing baseline trend, increasing baseline trend
- the the baseline data is already decreasing or increasing before the intervention it can be
confusing (ex if the baseline is decreasing and after the intervention it continues to decrease as
expected, did the intervention do anything? Or would the data have done that naturally? )
-Graphs determine IF change has occurred, not WHY is has occurred. Research can determine
- Research in applied behavior analysis typically uses single-subject designs: which examine the
target behavior of a person across time, while an intervention is either in effect or absent.
- an independent variable is tested for its potential or suspected influence
- a dependent variable is assessed to see if it’s value corresponds to, or “depends on” , variations
in the independent variable.
- When we examine why a behavior changed, we are seeking a cause-effect answer.
- We must control all extraneous variables, factors that could affect the dependent variable.
- When researchers find that a variable, such as reinforcement, causes a behavioral change, they
demonstrate a functional relation, that is the behavior changes as a function of the dependent
-AB design is the simplest type of single-subject research, consisting of one baseline phase and
one intervention phase
- The AB design may be useful when the purpose of research is simply to determine the extent to
which the behavior changed, but it is less than ideal when we want to isolate the cause of the
- Reversal designs: have a series of phases in which an intervention is alternately absent and
present, usually with either three or four phases.
- The ABA design: has three phases: baseline, intervention, and reversal. The last phase
withdrawals the intervention.
- The ABAB design: contains four phases: baseline, intervention, reversal (baseline),
intervention, reversal (baseline) and intervention.
Three problems: 1) effect of the intervention may not be f