Social%20Influence:%20A%20Neglected%20Issue%20in%20Social-Dilemma%20Research?

1
Social InfluenceA Neglected Issue in
Social-Dilemma Research?

• Tommy Gärling
• Maria Andersson
• Ted Martin Hedesström
• Department of Psychology
• University of Gothenburg, Göteborg Sweden

2
Social Influence
It is believed that sheep follow each other
mindlessly.
And men?
3
A Normal Day at the Stock Market

• In the stock market following others is referred
  to as herding. Conclusive market evidence of the
  existence and causes of herding is largely
  lacking.
• Experiments are conducted by asking
  participants to make choices in sequence such
  that in addition to private information each
  participant knows the choices made by the
  preceding participants.

4
Relationships to Social Dilemmas

• In social dilemmas under sequential protocol when
  participants know what others preceding them have
  contributed (taken), do they follow a majority of
  others due to their social influence? And what
  role does interdependence play in stock markets?

5
Theories of Social Influence

• Social influence creates conflict between the
  source and the target of influence.
• Two ways of solving the conflict include
• Conforming to a majoritys view Majority
  influence
• Accepting a minoritys view Minority influence
• Dual-process theories assume there are
  qualitative differences between majority and
  minority influences.

6
Theories of Social Influence (cont.)

• Majority influence is associated with heuristic
  processing
• (either because one wants to belong to the
  majority normative social influence or because
  one believes the majority is correct
  informative social influence)
• Minority influence is associated with systematic
  processing (critical assessment)
• Empirical studies (Erb, Bohner, Rank,
  Einwiller, 2002 Martin, Hewstone, Martin,
  2007 Martin, Martin, Smith, Hewstone, 2007
  Tormala, DeSensi, Petty, 2007) crossing source
  status and message processing give mixed results,
  some support dual-process theories and some do
  not. Several moderating factors.

7
Method Overview The Market
75
Upmarket mean 12
75
Downmarket mean 9
8
Method Overview Material (Upmarket)
Signal Participant A Participant B Participant C
1 gt10 Up Up Down
2 gt10 Up Down Up
3 gt10 Down Up Up
4 gt10 Up Down Down
5 gt10 Down Up Down
6 gt10 Down Down Up
7 gt10 Up Up Down
8 gt10 Up Down Up
9 gt10 Up Down Down
10 10 Down Up Up
11 10 Down Up Down
12 10 Down Down Up
9
Method Overview - Design

• Participants
• 60 undergraduates in different programs
• Design
• A 3 (payoff individual vs minority vs.
  majority)
• by 2 (market state up vs. down)
• by 5 (4) x 12 (trials) factorial design with
  trial as a repeated-measures factor.
• Payoff schemes
• Individual payoff If correct prediction 4 SEK
• (EV of following majority/minority 2 SEK,
  EV of following signal 3 SEK)
• Majority/minority payoff If correct prediction 2
  SEK 4 SEK if coinciding with the majoritys
  (minoritys) predictions
• (EV of following majority/minority 3/1
  SEK/EV of following signal 3 SEK)

10
Hypotheses

• Several previous experiments we have conducted
  (see Andersson, 2009) have shown that a majority
  has a larger impact than a minority. The payoff
  for individual performance (not following the
  majority) compared to the payoff for following
  the majority is expected to counteract following
  the majority if it instigates systematic
  processing (and following the majority leads to
  worse performance). The paypoff for following the
  minority may have no effect because a minority
  already instigates systematic processing.
• Andersson, M. (2009). Social influence in the
  stock market. Doctoral dissertation, University
  of Gothenburg, Göteborg, Sweden.

11
Results
Percent Correct
No significant effects (? .05)
12
Results (cont.)
Follow signal
Index
Follow herd
Main effect of payoff F(2,54) 4.46, p .016,
?2partial .06 Only the majority-payoff condition
differs reliably from the other conditions.
13
Results of Additional Experiments

• Individual payoff had a larger effect in the
  absence of the herd (percent correct 74 vs. 65)
• Only slightly less effect when majority payoff
  was eliminated (percent correct 69.7 vs. 60.6,
• index .36 vs. .17)

14
Discussion

• The results show that in the majority-payoff (and
  no bonus) condition(s) participants pay less
  attention to the signal and more attention to the
  others predictions than do participants in the
  other payoff conditions. An interpretation is
  that a majority is followed mindlessly
  (heuristic processing), whereas an individual
  payoff instigates systematic processing that lead
  to more accurate performance.
• In contrast a minority bonus had no effect
  because a minority already elicits systematic
  processing. (Not confirmed in consequent
  experiments.)
• In a social dilemma under sequential protocol,
  the interdependence certainly plays a role. Does
  social influence? If there is uncertainty and
  contributions are anonymous, informative social
  influence may dominate. If not anonymous
  normative social influence (reputation) would
  perhaps dominate.
• In the reverse, what does interdependence add to
  accounts of social influence (in stock markets)?
  

15
Thank you!
Questions and comments?
16
Social Influence
Almost mindlessly people follow others.
Anonymous thinker
17
Method overview - Task

• Task Predict upmarket or downmarket
• Information in all conditions
• - Before each trial, participants receive a
  signal.
• - The signal (ranging 5 to 16) is sampled from
  one of two distributions (upmarket or
  downmarket)
• - The predictions by 3 (fictitious) others

18
Method overview - Instructions

• Probability of initial market state
  upmarket/downmarket .50
• Maximum one change
• Change on any trial after the twelfth (no change
  occurred)
• Probability of market change .50