Title: Organization Culture as a Complex System: Balance and Information in Models of Influence and Selecti
1Organization Culture as a Complex System
Balance and Information in Models of Influence
and Selection
- Kenneth Frank, College of Education and Fisheries
and Wildlife - Kyle Fahrbach
2What Are Social Networks?
- A set of actors and the ties or relations among
them. - close colleagues (relation) among teachers
(actors) - help (tie) one teacher (actor) provides to
another - communication (tie) between people (actors) in an
organization - friendships (relation) among politicians (actors)
- links (relation) among web cites (actors)
- referrals (tie) among social service agencies
(actors)
3A Crystallized Sociogram
4A Ripple Plot
5What Makes a System Complex?
- Dual Models
- Influence
- Selection
- Dual Processes
- Information
- Balance
6Influence How Interactions Affect Beliefs and
Behaviors
- Research questions
- How does a teachers interactions affect her
implementation of innovations? - How does a bankers interactions affect her
profitability? - How does an adolescents interactions affect her
delinquency, alcohol use or engagement in school? -
- Theoretical Mechanisms
- Normative/conformity change to conform to
others around - Information change base don new information
- Dual processes both apply
7Basic Influence Model
kii indicates extent of relation between i and
i, as perceived by i. yit represents an
attitude or behavior of actor i at time t ?i
yit-1, the sum of the attributes of others to
whom actor i is related at t-1. Errors are
assumed iid normal, with mean zero and variance
(s2). Thus ? represents the extent of influence
of others on an actor And ? represents the extent
to which actors retain their own beliefs or
behaviors from one time to the next.
where
8Influence Graphical Representation
9Influence in Words (for teachers use of
computers)
use of computers time 2i ?use of first
colleague time 1 ?use of second colleague
time 1 ?use of last colleague time 1 ?(use
time 1)i error time 2i
10Model and Equation Toy Data
?KY1
?Y1
Y2
E2
0 1 1 0 0 0 1 0 1 0 0 0 0 1 0 1 0 1 0 0 0 0 1
1 0 0 0 1 0 0 0 0 1 1 0 0
2.4 2.6 1.1 -.5 -3 - 1
2 2 1 -.5 -2 -.5
2.4 2.6 1.1 -.5 -3 - 1
.11 -.01 .21 .13 .09 .13
0 1 0 1 0 1
x x x x x x
0 x 2.40 x (.15) 1 x 2.62.6 x (.15) 0 x 1.10
x (.15) 1 x -.5-.5 x (.15) 0 x -3 0 x (.15) 1 x
1-1 x (.15) Total 1.1 x (.15) .165
0 1 0 1 0 1
(.15)
(.67)
11For Actor 3
- y3 time 2 ?(y2 time 1 y4 time 1y6 time 1)/3
? y3 time 1 e3 time 1 - 1.15(2.6-.5-1)/3 .68(1.1) .21
12Influence Exercise
- Assume Bob talks to Sue with frequency 3, to Lisa
with frequency 2 and not at all to Jane. Last
year (at time 1), Sues delinquency behavior was
a 10, Lisas was a 5 and Janes was 2. - What is the mean of the influence of Bobs peers
regarding delinquency? - Hint ( Meansum/n, but what should n be?)
13Influence Model with Toy Data in SAS
- http//www.msu.edu/kenfrank/software.htmInfluenc
e_Models_ - influence program using proc means and merges
14Trajectories of Sentiment When all paths are
Operative
15Convergence of System to y1 y2
16Questions about W Row Normalization and
Interpretation of Influence
- Divide values by row marginal
- Different transformation for each subject
- Changes metric to influence units
- Access of one unit of expertise of one influence
unit increases number of uses of computers by xx
per year. - Theoretical meaning of influence units versus
frequency of interaction - Could you model influence unit with a selection
model?
17Redundant Effects through A Network
18Influence Model with All Paths
19Influence Model with Redundant paths
Where mpiit-v?t represents the maximal path from
experience of i at time t-v to I
(eit-v) Where qiit-v?t-mpiit-v?t represents
all paths other than the maximum. d governs
influence through redundant paths not
informational
20Trajectories of Sentiments When Effects of
Redundant Paths are Reduced
21Selection How Actors Choose Others with whom to
Interact
- Examples of Research Questions
- How do teachers decide to whom to provide help?
- How do bankers decide to whom to loan money?
- How do social service agencies choose other
agents to refer clients to? - Theoretical Mechanisms (references from Frank
Fahrbach 1999) - Balance seeking/homophily -- seeking to interact
with others like yourself - Information seeking Goal oriented
- Reduce uncertainty
- Power oriented
- Better understanding
- Curiosity
- Inoculate
-
- Evidence of Effects
22Basic Selection Model
In words Interaction between t-1 and t is a
function of kiit-t?t absolute value of
difference in previous beliefs a(k)yi t-1
yi t-1 prior tendency to
interact ?(k)kiit-t?t
23Selection Exercise
- 1) Write a model of whether two actors talked as
a function of whether they are the same race and
whether they are the same gender. - kii, represents whether i and i talked,
- yi represents the gender of i (0 if male, 1 if
female), and - zi represents the race of i (0 if white, 1 if
African American) - (Youll need one term for effects associated with
gender, and another for race)
24Selection Exercise
- 2) Assume Bob that and Lisa are African American
and that Jane and Bill are white. Bill and Bob
are Male and Lisa and Jane are female. -
- Calculate the independent variables based on
similarity of race and gender for Bob with each
of his interaction partners - (Bob, Lisa) same gender _______ same race
_________ - (Bob, Jane) same gender _______ same race
_________ - (Bob, Bill) same gender _______ same race
25Selection Exercise
- 3) Assuming the values of the ?s are positive
and that the effect of race is stronger than that
of gender, who is Bob most likely to talk to? - 4) Include a term capturing the interaction of
similarity of race and gender
26Model of Selection with Pursuit of new information
Where ? represents effect of pursuing maximal
paths
27Basic Model of Selection For Complex System
Where kiit-1?t represents interaction between
t-1 and t Yit-tit represents difference in
sentiments of actors a (k) represents homophily
birds of a feather ? (k) represents continuity
of interaction
28Oscillation Produced by Internal Dynamics of
System
29(No Transcript)
30(No Transcript)
31Conclusion
- Agent based modeling people are agents
- Influence
- Conformity
- Information
- People pursue (through selection)
- Balance (homophily birds of a feather)
- Information
- ? Changes in interaction as a result of influence
- What makes a complex system an organization?