Adolescent Romantic Couples

1
Adolescent Romantic Couples Interaction A
Cross-Study Analysis

• Joseph W. Dickson1
• Jill Carlivatii2
• Martin J. Ho3
• Deborah P. Welsh1

1 University of Tennessee 2 University of
Minnesota 3University of Denver
2
Background

• Collaboration brings the potential to advance
  research yet also has many challenges and
  barriers. Testing models using similar constructs
  in samples from diverse contexts can enhance the
  validity of findings.
• There has been a long and impressive literature
  examining the link between observed interactional
  processes of significant interpersonal
  relationships and the qualities of the
  relationship.
• These literatures suggest that relationships in
  which partners display fewer negative behaviors
  (such as conflict, negative reciprocity,
  hostility) and more positive behaviors (such as
  support, affection) report higher relationship
  qualities, longer lasting relationships, and
  healthier individual functioning.
• These literatures have also indicated that
  members of these relationships have very
  different understandings of their interactions
  within their relationship.

3
Aims

• Collaboration between three nationally funded
  longitudinal, multi-reporter, multi-method
  research projects
• Examine observed interactional processes and
  self-report of adolescent romantic couples
  aggression in predicting relationship
  satisfaction

4
Challenges

• Communication
• Conflicting schedules of busy researchers
• Different time zones
• Data management
• Unfamiliar with measures and data
• Finding similar constructs
• Personal Control Issues
• Allowing yourself to rely on others

5
Sample The data for this project come from three
independent projects.

• Site 1
• 211 adolescent dating couples1
• 93 couples between 17-21 yrs old
• Couples dating a minimum of 1 month
• (range 1-60 months mean 14.5 months)
• 1 The data for this project come from The Study
  of Tennessee Adolescent Romantic Relationships
  (STARR), funded by NICHD to Deborah P. Welsh,
  Ph.D.

6
Site 2

• 66 adolescent dating couples1
• 47 couples between 17-21 yrs old
• Couples dating a minimum of 6 months
• (range 6 to 47 months mean 18 months)
• 1A subset of participants from Studying Teens And
  Relationships (STAR), supported by Grant 50106
  from NIH to Wyndol Furman, Ph.D.

7
Site 3

• 86 adolescent dating couples1
• 72 couples approximately 21 yrs old
• Couples dating a minimum of 4 months
• (range months mean months)
• 1 Participants were a sub-sample of young adults
  taking part in the Minnesota Longitudinal Study
  of Parents and Children and supported by a grant
  from NIMH (MH 40964 -16) to Byron Egeland (PI),
  L. Alan Sroufe, and W. Andrew Collins.

8
Site 1 Site 2 Site 3
Region Southeastern U.S. Tennessee Western U.S. Colorado Midwestern U.S. Minnesota
Sample Size 93 Couples 47 Couples 72 Couples
Age 17-21 M18.3 17-24 M18.1 21
Weeks Dating 1 to 60 months M14.5 months 6 to 47 months M18 months 4 to xx months M months
Relationship Quality Self-report Relationship Satisfaction (Levesque, 1993) Self-report Quality of Marriage Index (Norton, 1983) Self-report Relationship Assessment Scale (Hendrick, 1988)
Self-Report Physical Aggression Modified Conflict Tactics Scale (CTS Straus, 1979) Modified Conflict Tactics Scale (CTS Straus, 1979) Modified Conflict Tactics Scale (CTS Straus, 1979)
Observed Interaction Couples Positive Connection Dyadic Positivity Shared Positive Affect
9
Relationship Satisfaction

• Site 1 Relationship Satisfaction (Levesque,
  1993) - a 5-item scale to assess satisfaction in
  adolescents romantic relationships. Participants
  were asked to responded using a six-point scale
  (1strongly disagree, 6strongly agree).
• Site 2 Quality of Marriage Index (QMI Norton,
  1983) - a six-item measure of an individuals
  global relationship satisfaction. Participants
  were asked to rate the degree to which each item
  characterized their romantic relationship on a
  seven-point Likert scale  (e.g., My relationship
  with my partner makes me happy.), with higher
  scores indicating greater relationship
  satisfaction. 
• Site 3 Relationship Assessment Scale (RAS
  Hendrick, 1988) - a 7-item scale of relationship
  satisfaction. Responses are given on 7-point
  scales (1low to 7high). The RAS has a coherent
  factor structure, is internally consistent.

10
Self-Report Aggression

• Modified Conflict Tactics Scale (Straus, 1979)
• Presence of Verbal Aggression (Self and Partner)
• Used insults
• Yelled
• Presence of Physical Aggression (Self and
  Partner)
• Pushed, grabbed, or shoved
• Tried to hit with something hard
• Hit with a fist

11
Interaction Procedure

• Site 1 Dyadic Positive Connection
• Participants recorded having a five-minute
  warm-up task and 2 eight-minute and forty-second
  interaction tasks (a conflictual issue selected
  by each couple member)
• Site 2 - Dyadic Positivity factor
• Participants recorded having 7 five-minutes tasks
  (a warm-up task, a problem selected by each
  couple member outside the relationship, a goal
  selected by each couple, a problem in the
  relationship, and a discussion of fun times)
• Site 3 Shared Positive Affect
• Participants recorded during two interaction
  tasks the Markman-Cox procedure (Cox, 1991).
  Couples decide on a topic causing most conflict
  and then discuss the problem for ten minutes
  trying to reach a solution.

12
Results

• Descriptive Findings Report of Verbal and
  Physical Aggression

Site 1 Site 2 Site 3
Verbal aggression (self) 51 96 91
Verbal aggression (partner) 51 92 85
Physical aggression (self) 18 25 11
Physical aggression (partner) 17 21 15
13
Results

• Descriptive Findings on Verbal Aggression

Site 1 Site 2 Site 3
Verbal aggression (neither partner self-reports) 29 0 6
Verbal aggression (one partner self-reports) 40 8.5 7
Verbal aggression (both partners self-report) 31 91.5 87
Verbal aggression (neither partner reports partner) 31 2.1 7
Verbal aggression (one partner reports other partner) 35 10.6 17
Verbal aggression (both partners report other partner) 33 87.2 76
14
Results

• Descriptive Findings on Physical Aggression

Site 1 Site 2 Site 3
Physical aggression (neither partner self-reports) 73 53 80
Physical aggression (one partner self-reports) 18 45 17
Physical aggression (both partners self-report) 9 2 3
Physical aggression (neither partner reports partner) 73 62 73
Physical aggression (one partner reports other partner) 20 34 24
Physical aggression (both partners report the other) 7 4 3
15
Results

• Relationship Satisfaction (HLM Baseline Model)
• Site 1
• 55 of variance predicting relationship
  satisfaction is due to individual
  characteristics
• 45 is due to couple differences
• Site 2
• 86 individual characteristics
• 14 couple differences
• Site 3
• 59 individual characteristics
• 41 couple differences

16
Results
Table 1 Multiple Regression Predicting
Relationship Satisfaction from Observers rating
of Couples Positive Affect and couple members
self-report of verbal and physical aggression.
Coefficients represent betas
Site 1 Site 2 Site 3
Observers rating of Positive Affect .94 -.10 .16
Verbal aggression (self) - -.28 -.11
Verbal aggression (partner) -.06 -.33 -.24
Physical aggression (self) - -.35 .12
Physical aggression (partner) -2.72 -.89 -.65
Variance explained (within/between couples) 8 of 55 38 of 45 8 of 86 7 of 59 21 of 41
p lt .05 p lt .01
17
Conclusion

• Communication is key
• Great opportunity to become more familiar and
  work with future colleagues and two world
  renowned projects .
• This project presents its own difficulties faced
  as an example of collaborative research
• Finding comparable data and constructs
• Although similar constructs were identified,
  other problems emerged.
• Differing within and between couples variability
  in outcome variable
• High rate of verbal aggression found in samples
• What predicts Relationship Satisfaction?
• Observers ratings of adolescent couples
  positiveness predicted relationship satisfaction
  in two of the three sites when controlling for
  couple members self-reported acts of aggression.
  
• Although not found in Site 2, this may be due to
  the amount of variance attributed to individual
  characteristics.