Chapter 12: Single-Subject Designs An alternative to experimental designsPurpose: To draw conclusions about the effects of

treatment based on the responses of a single patient under controlled conditions.

Based on: A research hypothesis indicating expected

relationship between independent and dependent variables

Specific operational definitions

Single-Subject Designs

Independent Variable-The intervention

Dependent Variable- The patient response (defined as the

target behavior) Target behavior is observable,

quantifiable, and a valid indicator of treatment effectiveness

Single-Subject Designs

Can be used to study comparisons between:

Several treatments Components of treatments Treatment and no-treatment conditions

Structure of Single-Subject Designs

Repeated Measurement Systematic collection of repeated

measurements of a behavioral response over time

These repeated assessment are required to observe trends or patterns and evaluate variability of the behavioral responses over time

Design Phases

Delineation of at least two testing periods:

• Baseline phase

• Intervention phase

• Target behavior is measured across both phases

Design Phases

Baseline information: Responses of target behavior during a period

of “no treatment” Reflects the target behavior over time in the

absence of the independent variable (intervention)

Changes from baseline to the intervention phase are attributed to the intervention

Design Phases

Design phases are plotted on a line graph

Magnitude of the target behavior along the Y-axis

Time (sessions, trial, days, weeks) along the X-axis

Baseline is represented by the letter A Intervention by the letter B

Design Phases

The design of one baseline period followed by one intervention period is: A- B design

Baseline data collection Unique to Single-Subject Design (all other designs treatment is initiated

following assessment)

Baseline Data Collection

Traditional designs make it impossible to determine:

Which component of treatment actually caused observed changes

If observed changes would have occurred without intervention

Baseline Data Collection

Baseline phase is a control period replacing a control group

Ethical considerations and baseline phase

Not unethical to withdraw treatment for a short period when we are not sure of effectiveness of treatment

Baseline Characteristics

Two characteristics of baseline data are important for interpretation of clinical outcomes:

• Stability- Consistency of response over time

• Trend- (slope) Shows the rate of change in the behavior

Baseline Characteristics

The most desirable baseline pattern demonstrates:– A constant level of behavior– Minimal variabilityIndicating: Target behavior is not

changingTherefore: Observable changes after

intervention are due to intervention

Baseline Characteristics

A variable baseline can present a problem for interpretation.

An Accelerating baseline-an increasing rate of response

A decelerating baseline-a decelerating rate of response

In both cases: a change in target behavior is occur13ring without intervention

Length of Phases

Flexibility in considerations depending on:– Type of patient– Type of treatment– Expected rate of change in the target

behavior

It is essential that the length of time within each phase is sufficient to capture any changes

Target Behavior

Can reflect:– Different response systemsMay focus on:

Impairmentsfunctional limitationsmeasures of disabilities

Measurements may deal with overt motor behaviors- functional performance, ROM, gait characteristics

Measuring Target Behavior

Frequency Duration Magnitude

Frequency

Counting the # of occurrences of the behavior within:

»A fixed time interval»Fixed number of trials»“Frequency count” is the simplest

of all behavioral measures

Frequency

Frequency count is appropriate to assess a discrete clinical behavior– Examples: – # of times a particular gait deviation occurs– # of times a client can repeat an exercise– # of times a patient loses her balance

during a treatment session

Frequency

Operational definitions for frequency counts must specify:– How the target behavior is distinguished

from other responses– What constitutes an occurrence and

nonoccurrence– (partial completion of exercise? fall over

but catching oneself?)

Frequency

“Frequency counts” are not useful when:– A behavior occurs too often to be counted

reliably– A behavior lasts for a long time (occurs too

seldom) The total time or total number of trials within

which the count is made must remain constant across sessions

Frequency

“Frequency counts” do not account for the quality of the behavior but only that it occurred

“Frequency counts” can be expressed as:– A percentage

• Dividing # of occurrences by total # of opportunities (percentage correct)

Frequency

Percentages are useful in that they are: Easily understood Efficient for summarizing large # of

responses Yet: If actual # of correct responses is

an indicant of the target behavior, percentage can be misleading

Frequency

“Frequency counts” can be translated into “rates”– The number of times a behavior occurs

within a specific time period (seconds, minutes, hours)

– Dividing the total # of occurrences by the total time

– (Ambulation in steps per minute)

Duration

Target behaviors can be measured according to how long they last

Duration can be measured either as:– The cumulative total duration of a behavior

during a treatment session– The duration of each individual

occurrences of the behavior

Duration

How long a patient stays in a balanced standing posture within:– A treatment session– Or:– Time how long it takes for a patient to

complete a functional task

Duration

Can be reported in terms of percentages

“Percentage time in zone”– (Dividing total time in the desired zone by

total time of training session)– This approach is useful when sessions are

not of equal length

Magnitude

Many clinical variables (target behaviors) are measured using instrumentation that provides quantitative data

(Electrical, functional performance)

Interval Recording for Observational Measures

Target behavior are usually recorded using either:– Quantitative instrumentation

• Appropriate for magnitude measure• Objective

– Self-report• Monitor activities outside the clinical

environment

– Direct observation

Interval Recording

Often recorded using frequency & duration methods to record the occurrence or nonoccurrence of the behavior

Certain behaviors are difficult to quantify– Break down the measurement period into preset

time intervals– Determine if behavior occur or does not occur

during each interval period (5 minutes)

Interval Recording

Sometimes called “time sampling” Total session time is divided into small

equal intervals Measurement may involve:

– Recording the presence/absence of the target behavior within each interval, and then tallying how many intervals contained the behavior

Interval recording

– Recording the frequency or duration of the behavior within each each interval

– It is important to select a time interval that will best reflect the expected frequency and duration of the behavior

– Requires the use of a signaling device

Reliability

Reliability is usually assessed concurrently with data collection, rather than in a separate pilot study

Reliability checks are performed by using two testers simultaneously observe the target behavior at several sessions across each phase

Reliability

Interrater reliability is usually reported using a measure of percentage agreement between observers

Total Reliability– Total steps: A=25; B=28; – Total reliability: (25/28)x 100= 89%– Limitation: Reflects only the consistency of the

total score for a session, but may observe different instances of the behavior

Reliability

Point-by-Point/Interval-by-Interval/Trial-by Trial

Agreement is based on: Number of occasions on which the observers agree that a behavior occurred or not occurred is divided by total occasions that raters agree and disagree

Total 30 trials observers agreed on 29: Trial-by-trial: (29/30) x 100= 97%

Reliability

Interval-by-interval– Of 16 intervals (15 minutes), observers

disagreed on 3 times (intervals 3,5,11)– (13/16)x 100= 81%– Chance agreement – Kappa – provides a statistical measure

Experimental Control

1. A-B: Baseline-Intervention (before-after)

2. A-B-A: Baseline-Intervention-Baseline (Withdrawal design)

If changes in behavior are not maintained during the second baseline phase- changes are due to intervention

3. A-B-A-B: In 3, 4 designs, behavior must be reversible

Experimental Control

Multiple Treatment Design1. A-B-C-B: Two treatments have independent

and differential effects2. A-B-A-C: A second baseline phase between

two treatments3. A-B-C-A-C-B: Sequential relationship

between B and C, and examine each treatment effect after baseline

4. A-B-C-BC: Combined phase

Data Analysis

Analysis is based on evaluation of measurements within and across design phases to determine if:

• Behaviors are changing• Observed changes during intervention are

associated with the onset of treatment

Data Analysis

1. Visual analysis– No mathematical operations– Intuitively meaningful– Data within a phase are described according to:

» Stability or variability» Trend- direction of change» Level- changes in magnitude (the value

of the behavior) from last data point of one phase to another

Data Analysis-Visual Analysis

Trend- direction of change within a phase• Accelerating or decelerating • Stable (constant) rate of change• Linear or curvilinear

A trend in baseline data:• No serious problem if against what is expected during

intervention• A slope of a trend can only be determined for linear data

Single-Subject Design

Now you know all about single-subject design