Single-Subject Designs

1
Single-Subject Designs
2
There are two broadly defined approaches to
experimental researchgroup designs single-subj
ect designs
3
Both approaches apply the components of the
scientific method to their approach to
research- the difference between group and
single-subject designs lies in the manner in
which the principles of the scientific method are
put into operation in designing studies
4
Single-subject designs are also called

• Time series
• functional analysis
• within-subject
• experimental analysis of behavior
• single-case experimental designs

5
Single-Subject Designs

• Use this design to demonstrate a functional
  relationship between changes in the IV and the
  resultant changes in the DV
• it provides an empirical verification that
  behavior change occurred because intervention
  occurred and no other cause

6
Components of Single-Subject Designs

• Single-subject designs are typically depicted by
  using letters A and B
• A the baseline and withdrawal phases
• B treatment phase
• there are different single-subject designs
• some are descriptive
• some are experimental

7
Descriptive Designs

• A baseline only (diary study)
• B treatment only (case study)
• AB baseline treatment (case study no control)

8
Experimental Single-Subject Designs

• Withdrawal Designs
• ABA
• BAB
• ABAB TREATMENT -
• Reversal Designs NO TREATMENT
• ABA COMPARISONS
• ABAB
• Multiple Baseline Designs
• MB x Behaviors
• MB x Subjects
• MB x Settings

9
Experimental Single-Subject Designs

• Interaction Designs
• Reduction (A - BC - B - C)
• Addition (A - B - C- BC)
• use these designs to look at interactive effects
  of individual components of tx
• Component assessment designs
• Alternating Treatments Designs
• B
• A - C - B or C or D
• D
• Treatment-Treatment Comparison

10
Components of Single-Subject Designs

• specify the problem
• select the design
• measure the problem
• repeated measures
• baseline
• analysis of data

11
Specify the problem

• Identify the behavior that needs to be changed or
  treated
• the behavior must be specified as clearly as
  possible in order for it to be reliably measured

12
Select the design

• There are a variety of SS designs that are
  defined by different phases
• select design on basis of question want to answer
• Withdrawal, Reversal or MB designs answer
  question Is treatment effective?
• Interaction designs answer question, What are
  the interactive components of treatment?
• ATD answer question, Which treatment is more
  effective?
• Phases in SS designs consist basically of a
  baseline (or no tx phase) and tx phases.
• These phases can be combined in different ways to
  derive different designs

13
Measure the problem

• After the behavior has been clearly defined, the
  clinician decides how the behavior will be
  measured,e.g., correct/incorrect responses, rate
  of response, length of response, etc.

14
Repeated Measures

• The behavior is repeatedly measured before,
  during, and after treatment to determine if any
  changes have occurred in that behavior.
• This component is the HALLMARK of single-subject
  designs

15
Baseline

• Before treatment is initiated, the behavior is
  measured over a period of time (1-2 weeks, few
  days)
• the BL provides a comparison of before and
  after treatment in order to evaluate the
  effectiveness of the treatment approach

16
Analysis of data

• Single-subject designs rely on visual analysis of
  the graphed data rather than on statistical
  analysis of the data to make decisions about the
  significance of the treatment approach
• slope, trend, and level of graphed data

17
Comparison of Single-Subject and Group Designs

• Group designs and SS designs share similar
  attitudes and components, but they use the
  components differently
• in SS designs, as in all experimental research,
  variables are manipulated and then you measure
  the effect of the IV on the DV
• SS designs differ from group designs in how they
  demonstrate experimental control

18
Comparison of Single-Subject and Group Designs

• Experimental control in SS designs is
  demonstrated in 2 ways
• in SS designs, the SUBJECT serves as their own
  control and is administered all conditions of the
  experimental condition and then all conditions
  are re-measured
• therefore, dont need a control group to
  demonstrate experimental control
• control behavior or control phase

19
Comparison of Single-Subject and Group Designs

• Group designs demonstrate experimental control in
  3 ways
• two groups (control group and experimental group)
• random selection of subjects from a larger
  population of subjects
• statistical tests
• help to look beyond the variability found in
  examining the individual performances of the
  subjects within each group.
• Thus, idiosyncratic behavioral patterns are
  ignored by pooling subjects data and the
  statistic serves as a control for extraneous
  variables

20
Comparison of Single-Subject and Group Designs

• Single-subject designs also differ from group
  designs in terms of how power/significance of the
  results is determined
• In SS designs, power/significance is demonstrated
  in 2 ways
• through the replication of tx effectiveness
  across AB phases within a subject and/or
  replication of tx effectiveness across subjects
• through the trend, slope, and level of the
  visually presented data

21
Comparison of Single-Subject and Group Designs

• In group designs, power is demonstrated through
  statistical significance

22
Summary Comparison of Group and Single-Subject
Designs Experimental Control

• Group Designs
• Control Experimental Groups
• Random selection of subjects
• Statistical tests

• Single-Subject Designs
• Each subject serves as OWN control
• Control behavior or control phase

23
Summary Comparison of Group and Single-Subject
Designs Power/Significance

• Group Designs
• Statistical significance

• Single-Subject Designs
• Replication
• across phases within subjects
• across subjects
• Visual inference of graphed data
• slope, trend, level

24
Visual Analysis of Single-Subject Designs

• Group designs determine the effectiveness/signific
  ance at predetermined probability levels, e.g.,
  .05 or .01

25
Visual Analysis of Single-Subject Designs

• Single-subject designs determine the
  effectiveness of the IV generally by visual
  examination of the graphed data
• therefore the reader of SS research must be
  cautious in interpreting tx effectiveness from
  visual inference
• 3 parameters important to visually presented data
  and graphs
• trend
• slope
• level

26
Trend

• The effectiveness of the IV is measured by
  comparing the direction of the behavior before tx
  and after tx is administered
• 3 directions the behavior may take
• increase in the occurrence of the behavior
• decrease in the occurrence of the behavior
• no change in occurrence of the behavior
• to be believable, the behavior must be measured a
  sufficient number of times to establish that it
  is not change before tx (min. of 3x recommended)
  and there must be a stable BL

27
Slope

• The degree of slope in the trend will indicate
  how strong the trend is.
• A pronounced slope is stronger evidence that the
  behavior is changing than if the slope is a
  gentle one

28
Level

• The level at which the behavior is occurring
  before treatment is important in determining the
  significance of the change in the behavior
  following tx

29
Criteria for Evaluating Single-Subject Designs

• Determine if
• the data are reliable
• the behavior was altered when intervention was
  administered (level)
• the change observed was significant (slope)
• the results are generalizable to other
  individuals
• the BL was stable (trend)
• there was no variability within phases and/or
  phases

30
Advantages of Single-Subject Designs

• Hard to find large N of subjects for group
  designs
• dont need a large N of allegedly homogeneous
  subjects with a particular communication
  disorder, random selection of Ss for sample and
  random assignment to groups
• ethical question to withhold tx for control group
• SS designs are clinically based - dont need to
  have a group of CD individuals who will not
  receive tx in order to have a control group

31
Advantages of Single-Subject Designs

• Get detailed information on individual subjects
  rather than pre-post scores
• you get individual data points across time and
  therefore can detect individual differences in
  response to tx
• easier to control for Ss since each Ss serves as
  his/her own control
• SS designs are intervention designs that can be
  used for exploring tx effectiveness which is a
  primary issue in CDIS

32
Advantages of Single-Subject Designs

• SS designs are practice based and practitioner
  oriented in that they are flexible and changes
  can be made during the course of tx if deemed
  necessary by the clinician
• Researcher-clinician gap is narrowed
• environment for conducting SS research
• can easily be incorporated into clinical practice
  without disruption
• conducted in clinic during clinic hours and
  regularly scheduled tx sessions

33
Advantages of Single-Subject Designs

• Provide clinicians with a tool for demonstrating
  accountability to their clients, administrators,
  and funding sources.

34
Disadvantage of Single-Subject Designs

• Reluctance to generalize results from single
  subjects to groups of individuals

35
Considerations in Using Single-Subject Designs

• Length of phases
• should be equal (each tx phase)
• however, this is not always possible
• therefore, usually use criterion levels and
  maximum number of sessions
• Counterbalancing
• whenever more than one behavior is treated
  sequentially, there is risk of order effects
• order effects can be controlled in MB studies by
  counterbalancing
• this requires a second subject who goes through
  study in an opposite training order from 1st Ss
• therefore, more than 1 Ss needed for MB studies

36
Considerations in Using Single-Subject Designs

• Selecting Behaviors
• 2 behaviors selected for tx in a MB study must be
  different (independent) enough that they must not
  impact each other
• PRINCIPLE 2 different behaviors and a single tx
• Baseline
• BL measures must be obtained on all DV responses
• the BL period must be sufficient enough to
  determine the stability of the response
• the BL must demonstrate a trend, slope, and level
  that is sufficiently different from the DV when
  the IV is applied.