Next topic: The External Validity of Laboratory Research

1

• Next topic The External Validity of Laboratory
  Research

2
Each person receives an outcome based in
part on his/her choice and in part on the choice
of the other.
Social Interdependence
At least 2 people
Each is free to choose between at least 2
options/behaviors.
It is common to represent the interdependent
relationship as a payoff matrix.
3
The Adam Smith Game So called after his notion
of the Great Invisible Hand
There is a no-brainer selfish (self-interested
choice) for each person. If everyone follows
their own self-interest, the result is the best
possible world Or, collective welfare is
maximized when everyone is selfish.
B is the best self interested choice for both
players. If both players choose B, the result is
10 points for each. There is no other outcome
where collective welfare is higher.
4
The Adam Smith Game
Easy question What of people do you think
choose B in this game? Practically everyone!
5
The Prisoners Dilemma Game
There is a no-brainer selfish (self-interested
choice) for each person. If everyone follows
their own self-interest, the result is the WORST
possible world Or, collective welfare is
minimized.
B is the best self interested choice for both
players. If both players choose B, the result is
9 points for each. There is no other outcome
where collective welfare is LOWER
6
Adam Smith and Prisoners Dilemma Side by Side
Note The same numbers appear in each game. What
makes the games so different is the relationships
between the numbers.
7
What of people choose A and B in each game?
A choosers around 60 B choosers around
40
A choosers around 2 B choosers around 98
These s are for the same group of people.
8
In 1968, David Messick and Charles McClintock
introduced the notion of Social Value Orientation
(SVO). As well see, SVO helps us understand
the choice s we just saw It also corresponds to
a personality trait.
9
SVO is a pair of attitudes/concerns/values. One
is the concern a person has with his/her own well
being. This component of SVO is called Self A
person may be positively self concernend (Self
), negatively self concerned (Self -) or
indifferent to the self (Self 0) The second
component of SVO is the persons concern with the
well being of others. This component of SVO is
called Other And, the person may be Other ,
Other or Other 0
10
Research in this area has focused on the
possibility of 5 different SVOs or 5 different
types of people Altruists Self 0, Other
I dont care what I get I just want the
other to get as much as possible.
(Alt) Cooperators Self , Other I want the 2
of us to get as much together as possible. (J
for joint gain) Individualists Self , Other 0 I
dont care what happens to the Other. I want to
get as much for myself as I can. (O for own
gain) Competitors Self , Other I want to
get as far ahead of the Other as possible, and
avoid falling behind him/her. (R, for relative
gain) Aggressors Self 0, Other (Agg) I
dont care what happens to me. I want the other
to get as little as possible.
11
Heres a pretty picture of the 5 possible SVOs
12
So, how about a theory of people based on their
Social Motivations?
13

• SVO is measured with the so-called decomposed
  game procedure developed by Messick and
  McClintock.
• You took the decomposed game measure in our
  last class meeting.
• A B C
• You Get 6 7 5
• Other Gets 5 3 0

An Altruist (Alt) would choose A (maximize
others points) A Cooperator (J) would also
choose A (maximize sum of points) An
Individualist (O)would choose B (get as much for
myself as possible) A Competitor (R) would choose
C (get as far ahead of the other as possible) An
Aggressor (Agg) would choose C (hurt the other as
much as possible)
14

• Using different decomposed games, you can
  distinguish between all 5 SVOs.
• Altruists are Alts,
• Cooperators are Js
• Individualists are Os
• Competitors are Rs
• Aggressors are Aggs
• A B C
• You Get 6 7 5
• Other Gets 5 3 0
• Alt O R
• J Agg
• A B C
• You Get 4 7 2
• Other Gets 5 3 0
• Alt J Agg
• O
• R

15

• So, what of Alts, Js, Os, Rs, and Aggs are
  there?
• Does sex matter? Are females more cooperative
  (J-ish) than males?

This distribution is seen in many countries,
including USA, Holland, Germany, Poland, Japan
and Korea.
16

• So, most people are either J, O, or R.
• Does this help us understand the choice
  percentages in Adam Smith, and Prisoners
  Dilemma?
• Yes.

17
Adam Smith Almost everyone chooses B.
A J would choose B. An O would choose B. An R
would choose B.
18

• Prisoners Dilemma
• Around 60 choose A
• Around 40 choose B

A J (around 55) would choose A An O (around 30)
and also an R (around 10) would choose B
19

• A really fair question at this point goes
  something like this
• These are only silly, imaginary games. How
  could they tell us anything important about
  social motivations in the real world?
• Or, do contrived laboratory tasks have any
  external validity?
• One great technique for addressing this question
  was developed by Daryl Bem, and is called
  Template Matching.

20

• Template Matching can be used to demonstrate
  external validity when
• Some lab task produces reliable individual
  differences.
• We know that the decomposed game is such a task.
• It suggests that people are Js, Os, Rs, at least
  in the lab.
• Do these differences extend to the real world?

21

• Procedures for Template Matching.
• Step 1 Generate Templates
• 1a Explain the individual differences procedure
  to a group of psychologists, (or, a group of
  students such as yourselves) so that when youre
  done they will know what is meant by Js, Os and
  Rs.
• 2a Give each psychologist (student) a Q-sort
  deck.
• This is a set of 100 cards, on which are
• written non-technical descriptions of the
  ways
• people act, think, and feel.
• For example One card says Likes to
  dominate conversations

22

• Each judge indicates the degree to which each
  statement sounds like, or is characteristic of
  a J, then an O, and then an R.
• For example Likes to dominate conversations
• How would you rate a J on this statement?
• 1 not at all characteristic
• 8 highly characteristic.
• How about an O?
• 1 not at all characteristic
• 8 highly characteristic.
• How about an R?

23

• After each SVO (J, O, R) has been rated on all of
  the Q-sort cards, the ratings are averaged, to
  produce what Bem calls templates (or profiles)
  for a J, for an O, and for an R.

24
Templates for Each SVO
These templates correspond to the theory held
by the judges as to the nature of J-ishness,
O-ishness, and R-ishness. Bem finds that the
judges tend to agree with one another. That is,
they appear to hold a common theory about each
SVO. This could only happen if the J, O, R
concepts had some a-priori meaning to the
judges. The question is this Does the real
world behavior of an actual J, or O, or R best
match the template for his/her corresponding SVO?
25

• This brings us to Step 2 of Bems procedure
  Template Matching
• Bring a student to the lab. Imagine its Joe.
• Measure his SVO. Imagine Joe is a J.
• Ask someone who knows this student well to come
  to the lab.
• Ask this acquaintance to rate Joe (the J) on the
  Q-sort deck.
• So, we now have a description of the real world
  behaviors of Joe, from an acquaintance who has no
  way of knowing why he/shes being ask to provide
  one.

26
Template Matching Which template is most similar
to Joes real world behavior?
This is a nice result! And it happened for 83 of
the cases in Bems study.
r 0.45
r 0.19
Similarity is measured by computing the
correlation coefficient between Joes profile and
each template.
r -0.24
27
Heres the scatter plot for the J-template, and
Joe the Js real world behavior (r0.45)
Why is the correlation so small? Why is the
scatter plot so messy? Well, we should EXPECT
it to be. And for reasons we talked about in an
earlier lecture . MULTIPLE DETERMINISM There is
more to Joes real world behavior than his
J-ishness.
28

• Does the situation have an influence on SVO?

29

• What happens if the other is a stranger, versus
  a friend?
• Recent research suggests that sometimes it
  doesnt matter, and other times it does.

30
Here is an Alt JORAg Game

• A B
• You Get 2 6
• Other Gets 7 4
• Alt
• Jnt
• Own
• Rel
• Agr

Column A Choices in This Game
31
Here is an Alt JO RAg Game

• A B
• You Get 7 5
• Other Gets 4 0
• Alt
• Jnt
• Own
• Rel
• Agr

Column A Choices in This Game
32
Here is an Alt JOR Ag Game

• A B
• You Get 7 5
• Other Gets 1 0
• Alt
• Jnt
• Own
• Rel
• Agr

Column A Choices in This Game
33

• So far, no difference between Friend and Stranger
  in ---
• Alt JORAg
• AltJO Rag
• AltJOR Ag
• Lets look at one more game.

34
Here is an AltJ OR Ag Game

• A B
• You Get 6 7
• Other Gets 4 2
• Alt
• Jnt
• Own
• Rel
• Agr

Column A Choices in This Game
35

• Whats the difference between these 3 games .
• Alt JORAg No effect for friend/stranger
• AltJO Rag No effect for friend/stranger
• AltJOR Ag No effect for friend/stranger
• And this one?
• AltJ ORAg Friend/Stanger Effect