Three Studies Testing a Model of Self-Reflexion

1
ABSTRACT
Three Studies Testing a Model of
Self-Reflexion Brenda L. McDaniel James W.
Grice Oklahoma State University
Three studies were conducted to assess Lefebvres
(1998) algebraic model of self-reflexion. In a
replication study, participants rated themselves
and eighteen others on fifteen bipolar adjectives
(e.g., excitable-calm). Half of the eighteen
individuals were positive (e.g., close friend)
and half were negative (e.g., disliked person).
The experimenters determined which adjective in
each pair was positive, and the following
proportions were predicted from Lefebvres model
(a) positive judgments of self, .719 (b)
positive judgments of others, .628 (c)
self-similar judgments, .619. The observed
proportions were .757, .606, and .615,
respectively. Ninety-five percent confidence
intervals indicated that only the like-self
proportion (.615 vs. .619) was not significantly
different from the hypothesized value. In the
second study, two additional sets of bipolar
adjectives were administered to different groups
of participants Big Five personality descriptors
(e.g., quiet-talkative) and unique personal
constructs. Participants were also asked to first
judge which adjective in each pair they
considered as more desirable before rating
themselves and the eighteen others. The following
proportions were predicted (a) positive
self-judgments, .844 (b) positive other
judgments, .613 (c) self-similar judgments,
.578. The observed proportions across groups
were, .851, .614, and .589, respectively.
Ninety-five percent CIs indicated that the
positive self (.851 vs. .844) and positive other
judgments (.614 vs. .613) were not significantly
different from the hypothesized values. The same
methods were used in the third study, except
participants rated themselves first before rating
the other individuals. The predicted proportions
were (a) positive self-judgments, .844 (b)
positive other judgments, .646 (c) like-self
judgments, .625. The observed proportions were
.857, .621, and .602, respectively. Ninety-five
percent CIs indicated that all were
significantly different from the predicted
proportions. The results from the three studies
were hence mixed, but encouraging. Further
experimentation and development of Lefebvres
model is recommended and will be discussed in
this presentation.    
Study 2
Discussion

•  Again, participants rated themselves and 18
  other individuals, in random order, on 15 bipolar
  semantic adjectives.
• Participants also rated themselves and 18 other
  individuals, in random order, on big five
  descriptors (e.g., quiet-talkative
  punctual-late calm-worrying) and unique personal
  constructs that each participant designed for
  themselves.
• Participants also judged which adjective in each
  pair was more desirable.

•  Mathematical modeling improvements from Study 1
  to Study 2 seem to have made predictions more
  accurate.
• In addition, it seems that it is important for
  the individual participant to choose which polar
  end of each adjective pair is more desirable or
  positive. This would be in line with George
  Kellys theory of personal constructs and how
  constructs must be personally relevant.
• Furthermore, it would seem that rating the self
  first had an effect on overall proportions.
  Perhaps thinking of the self first on all
  dimensions affect subsequent ratings and
  comparisons of others.
• However, interpretations of the present studies
  warrants caution.

REFLEXIVE STRUCTURE OF HUMAN COGNITION
RESULTS Study 2 (Results were collapsed across
groups)
b4image of the
other from others point of view

a4 image of the self from
others point of view

b3image of the other from selfs
point of view

a3 image of the self from selfs point of
view

b2 image of other

a2 image of the
self

a1The person in the
moment
a3
b3
a4
b4
a2
b2
a1
Future Directions
a4
a3
b4
b3

•  A true manipulation of the model is needed to
  come to more definitive conclusions.
• We are currently designing a study to manipulate
  social desirability, which is one of the tiers
  within the mathematical model.
• After this manipulation, more will be known about
  the models predictive power and the changes in
  structure that were implemented in an attempt for
  more accurate modeling of human cognition.
• The results presented are encouraging, not only
  to be able to make point predictions about how we
  structure our inner world, but for psychology as
  a science.

a2
b2
a1
Study 1

• An attempt to replicate findings of Lefebvre,
  Lefebvre, Adams-Webber (1986).
• Participants rated themselves and 18 other
  individuals, in random order, on 15 bipolar
  semantic adjectives (the positive adjective in
  each set was determined by the experimenter).
• Examples for others include A person in high
  school or middle school whom you did not like A
  person in high school or middle school whom you
  liked A current or past romantic partner whom
  you still love (or a person of the opposite sex
  whom you like).
• Examples for the bipolar adjectives include
  Generous-stingy strong-weak active-passive
  etc.

Note. Asterisks indicate proportions that are
significantly different from their respective,
predicted proportions (p lt .05, two-tailed).
Study 3

•  Study 3 was the same as Study 2 except
  participants rated themselves first, prior to
  rating the 18 other individuals.

RESULTS Study 1
RESULTS Study 3 (Results were collapsed across
groups)
Contact Information

•  A manuscript encompassing the present three
  studies is currently in progress. More details
  about the calculations and mathematical modeling
  will be presented. Furthermore, more details of
  the method will be given.
• Please send requests for further information
  about our research and questions/comments to
• Brenda McDaniel brenda.mcdaniel_at_okstate.edu
• Dr. James Grice jgrice_at_okstate.edu
• Thank you for your interest in our research!

Note. Asterisks indicate proportions that are
significantly different from their respective,
predicted proportions (p lt .05, two-tailed).
Note. Asterisks indicate proportions that are
significantly different from their respective,
predicted proportions (p lt .05, two-tailed).