SWL 579A Session 4SWL 579A Session 4

Instructor: J. David HawkinsUniversity of Washington

10/21/09

While you wait -While you wait -

What is the difference between sensitivity and specificity of a measure?

EPIDEMIOLOGY ETIOLOGY EFFICACY EFFECTIVENESS DISSEMINATION

1. Identify problem or disorder(s) and review information to determine its extent

2. With an emphasis on risk and protective factors, review relevant information-both from fields outside prevention and from existing preventive intervention research programs

3. Design, conduct, and analyze pilot studies and confirmatory and replication trials of the preventive intervention program

4. Design, conduct, and analyze large-scale field trials of the preventive intervention program

5. Facilitate large-scale implementation and ongoing evaluation of the preventive intervention program in the community.

The preventive intervention research cycle. Preventive intervention research is represented in boxes three and four. Note that although information from many different fields in health research, represented in the first and second boxes, is necessary to the cycle depicted here, it is the review of the information, rather than the original studies, that is considered to be part of the preventive intervention research cycle. Likewise, for the fifth box, it is the facilitation by the investigator of the shift from research project to community service program with ongoing evaluation, rather than the service program itself, that is part of the preventive intervention research cycle. Although only one feedback loop is represented here, the exchange of knowledge among researchers and between researchers and community practitioners occurs throughout the cycle.

Principles for Culturally Competent Prevention Science

• Include multiple groups in studies.• Compare prevalences and rates of

positive and problem behaviors and outcomes.

• Include measures of ethnic identify or acculturation to understand degree to which group differences reflect culture.

• Examine levels of risk and protection exposure in different groups.

Principles for Culturally Competent Prevention Science

• Examine the strength of association of risk and protective factors with outcomes in different groups after controlling for socioeconomic status.

• Use engagement approaches that effectively involve the focal audience.

• Use communication strategies that effectively communicate with the focal audience.

Perceived opportunities for prosocial

interaction/involvement with prosocial family,

classmates, teachers and classroom activities

Interaction/involvement with prosocial family, caregivers, teachers,

classmates, and classroom activities

Perceived rewards for interaction/involvement with prosocial family, caregivers, teachers,

classmates and classroom activities

Family management

Classroom management

School policies

Position in the social structure

Perceived opportunities for interaction with

antisocial family and caregivers, and/or

involvement in aggressive and other

problem behaviors

Interaction with antisocial family and caregivers, and/or

involvement in aggressive and other

problem behaviors

Perceived rewards for interaction with

antisocial family and caregivers, and/or

involvement in aggressive and other

problem behaviors

Skills for interaction/involvement

• Home-based services• Preparing for the Drug Free Years

• Respect & Responsibility• Proactive Family Management

• Newsletters

ANTISOCIAL PATH

PROSOCIAL PATH

(+) (-) (+,-)

Attachment and commitment to prosocial

family, caregivers, teachers, classmates

and classroom activities

Proactive ClassroomManagement

TeacherInstructional

Skills

Constitutional factorsInterpersonal and problem solving skills training and

summer camp

How to Help your Child Succeed in

SchoolAfter school study clubs

Belief in prosocial

family, caregivers, and school

values

Belief in antisocial family and caregivers’ values

Attachment and commitment to antisocial

family and caregivers

Drug and delinquency initiation

Note: Interaction or moderating effects are indicated by an arrowhead pointing to a structural path rather than a construct. Shaded circles indicate program interventions

The Social Development Model: Elementary School Period

Individual CharacteristicsBe Aware of…

The Social Development The Social Development StrategyStrategy

The Goal… Healthy Behaviors …for all children and youth

Healthy Beliefsand

Clear Standards

…in families, schools, and peer groups

Start with…

Build…Bonding

–Attachment–Commitment

…to families, schools, and peer groups

By providing… Opportunities Skills Recognition …in families, schools, and peer groups

Social development in a parent child interaction.

Parent-Child Interaction Coded for

OpportunitiesInvolvementRewardsBonding

etc.

Seattle Social Development Seattle Social Development ProjectProject

Targeted Risk FactorsTargeted Risk Factors School Domain

Low commitment to schoolAcademic failure

Family DomainPoor family managementFamily conflict

Individual DomainEarly antisocial behaviorFavorable attitudesFriends who engage in problem behaviorEarly initiation

Study CharacteristicsStudy Characteristics All 5th grade students from 18 Seattle

elementary schools were eligible. Active consent required – 77% (n=808) of

eligible population consented. Comparison Group Design

Full treatment (grades 1-6) = 149 Late treatment (grades 5-6) = 243 Control = 206 Parent training only = 208 (not included in these

analyses) Demographics

51% Male 46% Caucasian, 26% African American, 21% Asian-

American 56% free-lunch eligible 40% from single-parent families

Seattle Social Development Project: Hypothesized Seattle Social Development Project: Hypothesized Effects of Teaching on Students’ Social DevelopmentEffects of Teaching on Students’ Social Development

Student Opportunities For Classroom Involvement

Student Classroom Involvement

Reinforcement for Classroom Involvement

~ From Peers

~ Perceived Reinforcement for School Involvement

Bonding

~ To Prosocial Peers

~ To SchoolStudent Skills for Classroom Involvement

Teaching Interventions

Proactive Classroom Management

Interactive Teaching

Cooperative Learning

SSDP/SOAR Teaching SSDP/SOAR Teaching PracticesPracticesProactive classroom management

Establish consistent classroom expectations and routines at the beginning of

the yearGive clear, explicit instructions for behaviorRecognize and reward desirable student behavior and efforts to complyUse methods that keep minor classroom disruptions from interrupting

instructionInteractive teaching

Assess and activate foundation knowledge before teachingTeach to explicit learning objectivesModel skills to be learnedFrequently monitor student comprehension as material is presentedRe-teach material when necessary

Cooperative learningInvolve small teams of students of different ability levels and backgrounds as learning partnersProvide recognition to teams for academic improvement of individual

members over past performance

Proactive Classroom Proactive Classroom Management- Law of Least Management- Law of Least InterventionIntervention– Least amount of Least amount of

time time – Least amount of Least amount of

teacher effortteacher effort– Least amount of Least amount of

negative negative attention on attention on studentstudent

– Least Least unpleasant unpleasant feelingfeeling

– Least disruption Least disruption to the learning to the learning environmentenvironment

Eye contactEye contact ProximityProximity PausePause First/ThenFirst/Then Encouraging Encouraging

desirable desirable behaviorbehavior

Cueing Cueing HumorHumor EmpathyEmpathy Modify Modify

instructioninstruction

Comparison in Use of Project Comparison in Use of Project Teaching Practice Between Teaching Practice Between Intervention and Control Teachers Intervention and Control Teachers at Grade Sixat Grade Six

Low Implementers of Project Practices

High Implementers of Project Practices

Row Totals

Control Control TeachersTeachers

1365.0%

735.0%

2046.5%

Intervention Teachers

834.8%

1565.2%

2353.5%

Column Total

2148.8%

2221.2%

43100%

X2 = 3.9 p ≤ .05

Results of the Hierarchical Linear Modeling of Results of the Hierarchical Linear Modeling of Children’s Measures Nested within Teachers’Children’s Measures Nested within Teachers’

Hierarchical Linear Modeling

STUDENTCAT aggregate score

Classroom involvement opps.Skills School grades Social & academicClassroom involvementReinforcement From peers From schoolBonding To school To prosocial peers

X2 df p63.31 36 0.004

220.90 36 0.001

71.49 36 0.001148.51 36 0.001106.05 41 0.001

147.84 41 0.001244.12 41 0.001

222.65 41 0.00185.97 41 0.001

ONE-WAY ANOVAt p2.69 0.001

1.03 0.311

1.26 0.2151.24 0.205

-0.25 0.800

0.47 0.6420.47 0.638

0.27 0.786-0.29 0.774

INTERV. VS.CONTROL

t p1.53 0.134

3.14 0.004

1.11 0.275-0.86 0.3832.49 0.009

-0.11 0.9142.36 0.023

2.23 0.0320.70 0.478

IMPLEMENT.SCORE

Total Aggregated Risk Factors Total Aggregated Risk Factors PerspectivePerspective

J ohn A. Pollard , P h.D. Developmental Research and Programs

No studentsin this area.

Insuffic ient number ofstudents in this area.

Neighborhood #2

Neighborhood #1 Neighborhood #3

Madison Middle School Barriers to Learning Profi le 8th Grade

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Pe

rce

nt

At

Ris

k

School 2002 District 2002

Community Family School

Estimated National Value

Survey Participation Rate 2002: 87.4%

2002

Peer-Individual

Addressing Barriers with Effective Addressing Barriers with Effective ActionAction

3-5Early Childhood Education

prenatal-2Prenatal/Infancy Programs

6-14Family Therapy

prenatal-14Parent Training

Family Management Problems

Developmental PeriodProgram Strategy Factor Addressed

Effective Training for Middle School Effective Training for Middle School ParentsParents

Guiding Good Choices® (Spoth et al., 1998)Adolescent Transitions Program (Dishion and

Andrews, 1995)Parenting Adolescents Wisely (Gordon et al., 1998)Creating Lasting Connections (Johnson et al., 1996)The Iowa Strengthening Families Program (Spoth,

1998)Focus on Families (Catalano et al., 1999; 1997)

Nova High School Strengths Profi le 10th Grade

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Communityopportunities for

prosocialinvolvement

Communityrecognition for

prosocialinvolvement

FamilyAttachment

Familyopportunities for

prosocialinvolvement

Familyrecognition for

prosocialinvolvement

Schoolopportunities for

prosocialinvolvement

Schoolrecognition for

prosocialinvolvement

Social skills Belief in themoral order

OverallProtection

Pe

rce

nt

Pro

tect

ed

School 2002 District 2002

Community Family School

Estimated National Value

2002

Survey Participation Rate 2002: 79.7%

Peer-Individual Overall

Nova High School Substance Use & Antisocial Behavior

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Per

centa

ge

School 2002 MTF 2001

Ever Used 30-Day Use

Survey Participation Rate 2002: 79.7%

10th Grade: 2002HeavyUse

Past Year Antisocial Behavior

Nova High School Barriers to Learning 10th Grade

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Pe

rce

nt

At

Ris

k

School 2002 District 2002

Community Family School

Estimated National Value

Survey Participation Rate 2002: 79.7%

2002

Peer-Individual Peer-Individual

Addressing Barriers with Addressing Barriers with Effective ActionEffective Action

Risk Factor Addressed Program Strategy Developmental Period

Favorable Attitudes Toward Drug Use

Classroom Curricula for Social Competence and Competence Promotion

6-14

Community/School Policies

Classroom Curricula for Classroom Curricula for Social and Emotional Social and Emotional

Competence Promotion Competence Promotion Alcohol Misuse Prevention

(Maggs et al., 1998) Bicultural Competence Program

(Schinke et al., 1988)Towards No Drug Use (Dent et al.,

1995) The Valued Youth Partnership

(Cardenas et al., 1992)

Brown and Liao, Fig. 1. Three design phases of a preventive Brown and Liao, Fig. 1. Three design phases of a preventive intervention trial.intervention trial.

Pre-Intervention Assessment No Yes

Pre-Intervention Assessment No Yes

Target Population Pre-Randomized Design Not Located Contacted Or Contacted Ineligible Eligible Selection Bias Refused Consented Intervention Design Randomized Intervention Control Participation Participation Bias No Yes Post-Intervention Design Attrition

Sample / Selected No Yes

Intervention AssignmentIntervention Assignment

RandomizationRandomization Balance, Matching, BlockingBalance, Matching, Blocking Cluster Random Assignment Cluster Random Assignment

What are the Fatal Design What are the Fatal Design Flaws in a Trial?Flaws in a Trial?

Pre-Intervention Assignment:Pre-Intervention Assignment: Intervention:Intervention: Post-Intervention:Post-Intervention:

Pre-Intervention Pre-Intervention Assignment DesignAssignment Design

Extreme Selection BiasExtreme Selection Bias Not a Large Enough Sample is Not a Large Enough Sample is

DrawnDrawn

Intervention DesignIntervention Design

Intervention & Control Subjects Intervention & Control Subjects are differentare different

ContaminationContamination Randomized at Wrong LevelRandomized at Wrong Level Low Intervention DeliveryLow Intervention Delivery Large Drop-outsLarge Drop-outs

Post-Intervention DesignPost-Intervention Design

Large AttritionLarge Attrition Differential AttritionDifferential Attrition Differential Measurement ErrorDifferential Measurement Error

SSDP DesignSSDP Design

• Initiated full intervention and control conditions in 1981 in 8 Seattle elementary schools. • Expanded in 1985 to 18 Seattle elementary schools to add a late intervention condition, a parent training only condition, and additional control students.

• Quasi-experimental study

Full treatment (grades 1-6) = 149 Late treatment (grades 5-6) = 243 Control = 206

Parent training only (grades 5-6) = 210

No significant differences for those lost to attrition No significant differences for those lost to attrition versus those retained with respect to distribution of versus those retained with respect to distribution of participants into the intervention conditions at ages participants into the intervention conditions at ages 18, 21, 24 or 27.18, 21, 24 or 27.

No significant differences between the intervention No significant differences between the intervention and control groups with respect to:and control groups with respect to:• Gender, ethnicity, or childhood povertyGender, ethnicity, or childhood poverty• Mean years living in Seattle by grade 6 Mean years living in Seattle by grade 6 • Mean number of residences lived in from age 5 to 14 Mean number of residences lived in from age 5 to 14 • Proportion of single-parent homes during grade 5 Proportion of single-parent homes during grade 5 • Living in a disorganized neighborhood at age 16Living in a disorganized neighborhood at age 16• Family size, mother’s education, or age at time of survey at Family size, mother’s education, or age at time of survey at

age 21age 21BUT-BUT-More controls than full intervention group born to teen More controls than full intervention group born to teen

mothers.mothers.

Attrition and Internal Validity

2.70

2.75

2.80

2.85

2.90

2.95

3.00

3.05

3.10

13 14 15 16 17 18Age

Leve

l of Sc

hool

Bon

ding Full Treatment

Late TreatmentControl

Effects of SSDP Intervention Effects of SSDP Intervention on School Bonding from Age on School Bonding from Age

13 to 1813 to 18

Hawkins, Guo, Hill, Battin-Pearson & Abbott (2001)