TESTING THE LIMITS OF SOCIAL COGNITIVE THEORY: A cautionary tale for media effects research

1
TESTING THE LIMITS OF SOCIAL COGNITIVE THEORY A
cautionary tale for media effects research

• Robin L. Nabi
• University of California, Santa Barbara

2
WHATS THE PROBLEM?

• We say we value theory
• But how good are we at developing it?

3
TRENDS IN MEDIA EFFECTS RESEARCH

• Bryant Miron (2004) content analysis of 1,806
  articles in 3 leading journals between 1956-2000
• Only 26 theories/models mentioned more than 10
  times
• Only 5 mentioned more than 30
• Uses Gratifications
• Agenda setting
• Cultivation
• Social Learning
• Marxism

4
TRENDS IN MEDIA EFFECTS RESEARCH

• Potter Riddle (2007) content analysis
  2001-2006 (every issue of 16 mass media/comm
  journals)
• 8 theories cited more than 10 times
• Framing
• Agenda setting
• Cultivation
• Mediation models
• Third-person effect
• Uses and gratifications
• Selective exposure
• Social cognitive/learning

5
RAISES THE QUESTIONS.

• Why these theories?
• Are those good reasons?
• Could we do better?
• If so, how?

6
WHY THESE THEORIES?

• WIDELY APPLICABLE IN IMPORTANT CONTEXTS
• (Framing, Cultivation, Uses Gratifications)
• EASY TO UNDERSTAND, TEST, APPLY
• (Cultivation, Uses Grats)
• WILL SURELY FIND EFFECTS
• (3rd Person, Cultivation, Uses Grats)
• COMPELLING EVIDENCE IN OTHER, NON-MEDIA CONTEXTS
• (Social Cognitive Theory)

7
ARE THESE GOOD REASONS?

• DEPENDS ON GOALS
• Helps ID effects/confirm theoretical predictions
• More likely to get published
• More appeal to wider-audience
• But as media scholars. Raises concerns.

8
REASONS FOR CONCERN

• THEORETICAL STAGNATION
• Content analyses shows
• 48 (35 since 2000) simply reference theory
• 26 used theory as a framework
• Note Not necessarily used correctly
• 26 offered theoretical critique, comparison, or
  development
• Note yay! Is it productive critique?

9
REASONS FOR CONCERN

• DESCRIPTIVE RESEARCH ? MINIMAL INSIGHTS INTO
  PROCESS
• E.G. Uses Gratifications
• FAILURE TO OFFER SUITABLE TESTS
• E.G., Framing, Cultivation, SCT
• BLIND APPLICATIONS ? POTENTIAL MISAPPLICATIONS
• Social Cognitive Theory

10
SCT BRIEF OVERVIEW

• Bandura (1977, 1986, 2002)
• Focus on OBSERVATIONAL LEARNING SELF-EFFICACY
• Building from capacity for abstract thought
  goal setting

11
SCT 4 KEY PROCESSES

• 4 PROCESSES
• ATTEND to certain models (e.g., like, attractive)
• RETAIN symbolic representation of behavior and
  consequences
• PRODUCE behavior in appropriate context
• MOTIVATED to produce those selected behaviors
  based on nature of reinforcement
• IN ESSENCE show liked models engaged in healthy
  behaviors (pos reinforced) ? audience more likely
  to engage in those behaviors

12
THE PROBLEM

• SCT frequently cited to guide depictions of risky
  behavior in media
• BUT rarely (well) tested in media
• Cause to question the appropriateness of its
  application in media contexts

13
ALTERNATIVE VIEW Schema

• Schema Serial TV Programming
• Main characters dont experience truly bad,
  long-lived outcomes (Happily ever after)
• This future expectation influences how viewers
  interpret current story events
• Minimizes negative valence of negatively-reinforce
  d behaviors
• Increase positive valence of negatively-reinforc
  ed behaviors because liked characters do them

14
STUDY 1 PROGRAM SCHEMA (Nabi Clark,
under review)

• H1 TV viewers believe main characters in
  fictional TV series will experience positive
  outcomes, despite the adversity they face.
• RQ1 What explanations do TV viewers offer for
  expecting positive outcomes?

15
STUDY 1 METHOD

• 60 Undergraduates
• Survey (after finale of 2007 TV season)
• 6 closed-ended items
• Likelihood M.C. suffer long-lived consequences,
  happy ending, etc.
• 7 closed and open-ended items
• Likely outcomes of specific cliffhangers (e.g.,
  Desperate Housewives, CSI, Greys Anatomy, Lost)
• Why? (coded into 7 categories)
• bad things wont happen to the M.C
• contract disputes
• secondary characters are disposable

16
STUDY 1 RESULTS Expectations of Positive and
Negative Outcomes for Main Characters in TV
Serial Programming
17
STUDY 1 Specific Program Outcome Expectations
18
RESULTS RQ1Open-Ended Responses
19
STUDY 1 RESULTS

• Across 7 cliffhangers, 41 said expected outcome
  they did because main character is important to
  the program
• 100 gave this answer at least once
• THUS, American viewers TV schema Happily Ever
  After

20
STUDY 2 COMPETING TEST OF SCT v. SCHEMA

• H2 TV depiction should influence those without
  direct experience, but not those with direct
  experience.
• AMONG THOSE WITHOUT DIRECT EXPERIENCE
• H3 Neg. Depiction lt Positive Depiction
• (SCT Prediction)
• H3ALT Neg. Depiction Positive Depiction
• (Schema prediction)

21
STUDY 2 METHOD

• 400 women completed survey re past sexual
  behavior (valid n 238)
• Focus on One Night Stands
• 30 of sample had personal experience
• Viewed 1 of 6 edited versions of Sex The City
  episodes depicting One Night Stands
• 3 manipulations of rewards punishments
• Regret vs. no regret
• Lesson learned vs. acceptance of behavior
• Negative vs. positive outcomes

22
EXAMPLE OF STIMULI
23
STUDY 2 METHOD (cont.)

• POSTTEST (Key measures)
• Enjoyable (a .93)
• Carrie Likeable (a .84)
• ACCEPTING OF CARRIES BEHAVIOR (a .83)
• HOW REGRETFUL CARRIE SEEMED
• FUTURE LIKELIHOOD OF ONS
• Attitude toward friend who had a ONS (a .72)
• Homophily (a .80)
• Seen episode before

24
PRELIMINARY ANALYSES

• No differences across conditions in any pre-test
  measure
• No differences in enjoyment, similarity,
  likeability, etc.

25
MANIPULATION CHECKS

• REGRET
• R gt NR, p .02
• LL gt A, p .001
• NEGATIVE VIEW OF BEHAVIOR
• N gt P, p .04

26
RESULTS H2 (supported)

• Those without direct experience evidenced
  positive change. Those with direct experience
  did not.

27
H3 SCT v. SCHEMA

• 3 ANCOVAs
• IVs ONS experience, Viewing condition
• DV post likelihood of ONS
• Covariate initial likelihood of ONS
• MAIN EFFECT FOR EXPERIENCE supports H2
• INTERACTION (in expected direction) supports H3
• MAIN EFFECT, BUT NO INTERACTION supports H3alt

28
RESULTS H3

• NO MAIN EFFECTS for different depictions (ps
  .22 - .72)
• MAIN EFFECTS FOR CONDITIONS
• R/NR p .007
• LL/A p .01
• N/P p .06
• NO INTERACTIONS between previous experience
  depictions
• LL/A p .27
• N/P p .23
• R/NR p .03 BUT in the wrong direction (those
  with direct experienced were influenced, not
  those without)
• THUS, regardless of depictions, those without
  past experience reported increased likelihood re
  future ONS

29
CONCLUSIONS

• Results support Schema over SCT explanation
• Should not think of depictions in isolation but
  in the cognitive context in which received
• SCT perhaps more useful for depicting positive
  health behaviors than warning against negative
  ones
• Emphasize importance of carefully testing
  interpersonally-based theories in media contexts

30
FUTURE DIRECTIONS

• Perhaps examine harsher negative outcomes
• Consider more carefully role of identification
• Test other elements of SCT
• Self-efficacy
• Perceptions of what constitutes negative outcomes

31
MORAL TO CAUTIONARY TALE?

• MUST NOT BLINDLY ACCEPT ESTABLISHED THEORIES AS
  LAW
• CHALLENGE ASSUMPTIONS
• Look to other disciplines
• Consider unique features of media contexts
• OFFER SERIOUS TESTS
• OFFER MULTIPLE TESTS
• OFFER COMPETING TESTS
• DONT BE AFRAID OF HARD RESEARCH
• DONT THROW BABY OUT WITH THE BATH WATER

32
GLASS IS HALF FULL

• PROMISING LINES OF RESEARCH
• 3rd person research
• psychological explanations
• Framing research
• moderators
• Cultivation
• stagnated since Shrum
• UG
• developments in mood management
• SCT
• some newer studies offering more direct tests

33
FUTURE OF MEDIA EFFECT THEORIZING?

• LOTS TO DO ITS UP TO YOU!
• ICA SPOTLIGHT PANEL
• READ WIDELY
• TAKE YOUR TIME
• QUALITY OVER QUANTITY
• MAKE IT MEANINGFUL