Single Case Design Overview and Examples of SCD

Single Case Design: Overview and Examples of SCD in Home Visiting Evaluation Kate Lyon, MA James Bell Associates, Inc. Julie Morales, Ph. D James Bell Associates, Inc. David Gast, Ph. D University of Georgia April 2013

Webinar Overview • Today’s webinar: – Overview of single case design basics – Examples of home visiting program evaluation questions & designs 2

Single Case Design: • Rigorous • Tests impact on a single case • Determines general impact using small sample size • Visual analysis of data • Each case serves as its own control 3

Dependent Variable = observable target behavior that should be changed by the intervention • Precisely define target behavior to allow for measurement of the relative occurrence of the behavior before and after an intervention 4

Your definition of a dependent variable should be: ü Objective ü Clear ü Complete Two Examples: 1. Increase positive parenting practices OR 2. Increase parents’ rates of verbal praise for children and decrease rates of verbal criticism 5

Dennis’ definitions as a prevention scientist… Behavior = Reliably measurable actions (frequency, duration and/or intensity) of a person (adult or child) in a specific environment. 6

Frequency = counting how many behaviors in a certain time (e. g. , per minute, per hour) • What behaviors on your list might you count by frequency when at a home visit? 7

Duration = how long a specific behavior lasts (e. g. , seconds, minutes, hours) • What behaviors on your list might you count by duration at a home visit? 8

Intensity = positive or negative strength (i. e. , voice, gestures, expressions) • What behaviors on your list might you count by intensity at a home visit? 9

Dependent variable is measured repeatedly to allow for identification of behavior patterns Measures need to be sensitive to change Measurement of dependent variable must be monitored for consistency by measuring inter-observer agreement 10

Establish a Stable Baseline 7 6 5 4 3 2 1 0 1 2 3 4 5 7 stable or instable ? 6 5 4 3 2 1 11 0 1 2 3 4 5

What are your tools (not program) to increase or decrease the behaviors on your list at a home visit while you are there? 12

Independent variable = the practice, treatment or intervention being tested that should produce observable changes in the dependent variable • Is it time limited? • Can it be started and stopped? • Why will the IV change DV? 13

Single Case Design Home Visiting Program Evaluation Examples 14

Example #2: Parent-Child Interaction Single Case Design Do mothers who receive HV model components focused on parent-child interaction demonstrate improved parenting interaction with their children compared to their interaction prior to receiving the HV model components? PICO P = Population: Mothers with children ages 2 -5 years participating in the home visiting program I = Intervention: Selected home visiting model curriculum components focused on parent-child interaction C = Comparison: Mothers’ interaction with children prior to receiving HV model components focused on parent-child interaction O = Outcomes: Positive parent-child interaction behaviors as measured by an observational tool 15

Example #2: Parent-child Interaction Single Case Design • Dependent Variable = observable target behavior that should be changed by the intervention • Precisely define target behavior: Parent-child interaction behavior • Select appropriate instrument: reliability, validity, sensitivity of observational measure • Create schedule of observations for each phase that involves direct and frequent measurement – Baseline, intervention, effect: how many over what period (e. g. , 1 observation per week for 1 month) • Establish stable baseline observations prior to introduction of HV program components 16

Example #2: Parent-Child Interaction Single Case Design • Independent Variable: the practice, treatment or intervention being tested that should produce observable changes in the dependent variable • Examine the specific HV model component that is focused on improving parent-child interaction – Identify key content to be able to monitor that delivery of content occurs according to plan • The application of the specific curriculum component is observed and measured over time 17

Example #2: Parent-Child Interaction Single Case Design • Selecting your sample: sample participants must be randomly selected from all eligible participants • Select 3 to 5 participants • Sample participants must be randomly assigned to staggered intervention start points 18

Example #2: Parent-Child Interaction Single Case Design: Multiple Baseline Design HV Pgm Client #1 O O O baseline HV Pgm Client #2 O O O intervention O O O baseline HV Pgm Client #3 O O O effect O O intervention O O baseline O O O effect O O O intervention O O effect O

Example #2: Parent-Child Interaction Single Case Design • Analyzing the results: graphing changes • Provides a detailed visual description and numerical summary of behavior • Allows analysis of the relation between independent and dependent variables • Provides one visual display: – – – Sequence of conditions and phases Time spent in each condition Independent and dependent variables Evaluation or research design Relations between variables Spriggs and Gast (2010). Visual representation of data. In D. Gast (Ed. ) Single Subject 20 Research Methodology in Behavioral Sciences (pp. 166 -198). New York: Routledge.

Example #2: Parent-Child Interaction Single Case Design • Assessing effects: examination of data patterns within and across phases – – – Level (mean within a phase) Trend (slope of data points within a phase) Variability (range of data points around slope) Immediacy of effect (rapid effect between phases) Overlap (of data points between phases) Consistency (data patterns across similar phases) For more complete information see “What Works Clearinghouse”: http: //ies. ed. gov/ncee/wwc/pdf/wwc_scd. pdf 21

Questions? 22