Lecture 18: Single-Subject Designs Flashcards
how do we statistically analyze between-subjects designs?
an independent subjects t-test or between-subjects ANOVA
how do we statistically analyze within-subjects designs?
a dependent subjects t-test or repeated measures ANOVA
Single-subject (single-case) designs
research designs that use the results from a single participant or subject
goal of single-subject designs
Establish the existence of cause-and-effect relationships
how are single-case experimental designs conducted
by manipulating an IV and controlling extraneous variables to prevent alternative explanations for the research results
alternative explanations for research results in a single-subject design
- Baseline
- Repeated observations
- Replication
evaluating results from a single-case study
- Data are evaluated in a simple graph because statistical tests for significance cannot be used
- Contrary to what is stated in the textbook, several statistical significance tests could be used to compare the level and trend between the A and B phases
- However, such a graph by itself is not enough to show that the treatment caused a behavioural change
why are graphs not enough to show that the treament caused a behavioural change?
- No control over extraneous variables
- This could be the result of chance
phase
series of observations of the same individual under the same conditions
when are baseline observations made?
when no treatment is being administered
baseline phase
a series of baseline observations
baseline notation
Identified by the letter A
treatment observations
observations made when a treatment is being administered
treatment phase
a series of treatment observations
treatment notation
- Identified by the letter B
- Subsequent treatments are identified by subsequent letters (ex. c, d, e)
function of baseline and treatment annotation
Allows researchers to describe the phases in a study by using a sequence of letters
consistent level
- A series of measurements that are all nearly the same magnitude
- Graphed data points cluster around a horizontal line
consistent trend
- Differences from one measurement to the next that are consistently in the same direction and nearly the same magnitude
- Graphed data points cluster around a sloping line
stability
- The degree to which the observations show a pattern of consistent level or consistent trend
- Stable data may show minor variations from a perfectly consistent pattern
- Variations should be relatively small and the linear pattern relatively clear
strategies for dealing with unstable data
- Wait until the data stabilize
- Average a set of two or more observations
- Look for patterns within the inconsistency
length of a phase
- To establish a pattern (level or trend) within a phase and to determine the stability of the data within a phase, a phase must consist of a minimum of three observations
- Two observations, by themselves, do not provide enough information to determine a pattern.
- Additional observations beyond the first two are essential to establish level, trend, and stability.
- Typically, five or six observations are necessary to determine a clear pattern
- When high variability exists in the data points, additional observations should be made
phase change
a manipulation of the independent variable
what are accomplished with phase changes?
- Implementing a new treatment
- Withdrawing a treatment
- Changing a treatment
function of changing phases
Initiates a new phase in which the researcher collects observations under new conditions
when should we change phases?
When a clear pattern has emerged from the preceding phase
what factors should we consider when assessing participants’ responses to phases?
- If they are improving without treatment, don’t implement the treatment.
- If they are deteriorating quickly, implement the treatment right away.
- If the treatment produces immediate and severe deterioration, cease or modify treatment immediately.
visual inspection techniques
characteristics that help determine whether there is a meaningful change between phases
examples of visual inspection techniques
- Change in the average level
- Immediate change in level
- Change in trend
- Latency of change
tools to enhance visual inspection techniques
- Use lines representing a level in each phase and a band representing +/- 2 standard deviations
- Use lines representing the trend in each phase
phases of an ABAB design
A baseline phase (A), followed by treatment (B), then a return to baseline (A), and finally a repetition of the treatment phase (B)
goal of an ABAB design
to demonstrate that the treatment causes changes in the participant’s behaviour
demonstrating that the ABAB design causes a change in behaviour
- The pattern of behaviour in each treatment phase (different from the pattern in each baseline phase)
- The changes in behaviour from baseline to treatment and from treatment to baseline (the same for each of the phase-change points in the experiment)
limitations of the ABAB design
- Withdrawal from treatment
- This may not result in a change in behaviour (the patient is cured)
- This may result in a slight change but not a return to the baseline
- This may cause an ethical problem if the treatment is working for the participant
variations on the ABAB design
- Sometimes researchers add a treatment or modify the sequence of baseline and treatment phases.
- Creates a complex design
- Researchers must incorporate new treatments into the phase sequence to prove a causal relationship.
- ABBCAC, ABCBC, etc.
multiple baseline design
- The treatment phase is initiated for one baseline
- Baseline observations continue for the other
- The treatment is initiated for the second baseline at a different time
multiple-baseline across subjects
the initial baseline phases correspond to the same behaviour for two separate participants
multiple-baseline across behaviours
the initial baseline phases correspond to two separate behaviours for the same participant
multiple-baseline across situations
the initial baseline phases correspond to the same behaviour in two separate situations
component analysis designs
- Each phase adds or subtracts one component of a complex treatment
- Determines how each component contributes to the overall treatment effectiveness
- Component analysis with a reversal design
- Component analysis with a multiple-baseline design
rationale for multiple-baseline designs
- The criteria for a successful multiple-baseline experiment are essentially identical to the criteria described earlier to define the success of an ABAB design
- There is a clear and immediate change in the pattern of behaviour when the researcher switches from a baseline to a treatment phase
- The design includes at least two demonstrations that behaviour changes when the treatment is introduced
- This replication is necessary to establish a causal relationship between treatment and behaviour
strength of a multiple-baseline design
- No return to baseline needed
- Good for long-lasting treatments
weaknesses of a multiple-baseline design
- It can be difficult to identify similar but independent behaviours
- Results can be compromised by individual differences between participants or between behaviours
strengths of single-case designs
- Allows researchers to establish cause-and-effect relationships with one participant or subject
- Flexibility: the researcher is free to modify the treatment or change to a new treatment if a participant or subject fails to respond to the treatment
- No need to standardize treatment across groups—a single participant or subject is used
weakenesses of single-case designs
- Relationship among variables is for only one participant or subject
- May threaten external validity (generalization)
- Multiple, continuous observations are required
- Absence of statistical controls
- Reliance on graphs to display data
- Treatment effects must be large and immediate to produce a convincing graph