The Consequences of
Guilt by Association as an
Investigative Decision Rule
Elise A. Weaver,
dios Kent Rissmiller
Social Science and Policy Studies
Worcester Polytechnic Institute
&
George Richardson
Public Administration & Policy
University at Albany, SUNY
Nelson A.
Rockefeller
Oleg contact email:
of Public Affairs .
nd eweaver@wpi.edu
and Policy
Background
Substantive Insights
- Cole, David & Dempsey, James X.
(2002). Terrorism and the Constitution:
Sacrificing Civil Liberties in the Name of
National Security. NY: The New Press.
Background
Psychological Insight
_ Hammond, K.R. (1996).Human Judgment
and Social Policy: Irreducible
Uncertainty, Inevitable Error,
Unavoidable Injustice. NY: Oxford
University Press
Background
Initial System Dynamics Model
_ Weaver, E.A. & Richardson, G.P.
(2002).Threshold setting and the cycling
of a decision threshold. Presentation to
the 20th Conference of the System
Dynamics Society, Palermo, Italy, July
23, 2002.
Definitions of Key Terms
¢ Signs of criminal activity
oP — criminal activities or plans by specific
| individuals
— individuals sending money earmarked
for criminal activity
Definitions of Key Terms
¢ Guilt by association
— recent contacts of the guilty
— friends of the guilty
— those financing humanitarian causes
conducted by same organization as the
guilty
— those sharing ideology of guilty
— those sharing ethnicity or race of guilty
Key Components
¢ Judgment rule:
— Relative weight on signs of individual
criminal activity vs. guilt by association
¢ Quality of judgment
— Sensitivity = guilty detected/ all guilty
— Specificity = innocents left alone/
all innocents
Key Components
¢« Stakeholder pressures:
— Lobby to increase investigative power
— Protests as check on unrestrained
investigative power
¢ Outlay of Resources:
— Cost effectiveness of investigation
— Results of congressional audits
Key Components
¢ Legal Issues:
— Secrecy and breaches of due process
— Resulting lawsuits
¢ Public Response
— Community support for investigation
— Chill or decreased sense of freedom of
expression
— Tolerance for security measures such
as wait times at the airport
Highly Effective Judgment Rule
Overlook Individual Search Individual
Actual Guilt
Truly} Truly
Innocent} Guilty
Perceived Guilt
Investigative
Threshold
Less Effective Judgment Rule
Overlook Individual | Search Individual
Actual Guilt
Truly} Truly
Innocent} Guilty
Perceived Guilt
Investigative
Threshold
Reference Modes
¢ Oscillation of investigative power
over time
¢ High weight on Guilt by Association
resulting in exponential growth in
resources required
¢ Constraints on protest and secrecy
of investigation leading to
exponential growth of resources
required and poor judgment efficacy
WPI
Regulation Sector
Be Community Tolerance — f )
for Security Measures Specificity = B
r _ \ Innocents Left Alone/«—
ry Civil Liberties « - Actual Innocents
ae. Protest:
oe ( "Chill" on Freedom of 2
j 5 Expression D
Net Power Weight on
Change Power to
investigate n Guilt by
eS ead without ~~ Association
~ t+ regulation Relative Power
Security Normal Power RB
Conscious Lobby Sensitivity =
F Audits by ~ Guilty Detected/
| Congress a Actual Guil
WPI
Resource Allocation Sector
Contacts
Resources Spent Investigated per
a’ to Investigate | _A Person
<Weight on R Contacts
Guilt by YN Resources Requested
Association> SS qa B
Y +
\ Resources 4-9
————ooea to. Resources Allocated
a Weight on Signs ww by “
_ of Criminal Resources Spent
Activity to Gather Signs \. B
* NK of Individual Audits by
Criminal Activity Congress
Investigation of Contacts
<Contacts Fraction Perceived
Investigated per Innocent
Person> )
Number
Number of Perceived
Investigated
that are
Investigated
R Fraction
Perceived Guilty
Number
Perceived Guilty
that are
Investigated
Community Response
<Number of
Innocents =
Investigated>
\; Willingness of R ;
Public to Share 6 price P
Information amere
<Contacts 7
Investigated ver 7 Individual
Person> Requirement ¥ Criminal
for Secrecy WA Activity
<Resources Spent to Gather
Signs of Individual Criminal
Activity>
Legal Oversight
Specificity =
Innocents Left Alone /
Actual Innocents
B
- [
Number of
Lawsuits
- Ne
"Chill" on Breaches of
Freedom of Due Process
Expression «Requirement
for Secrecy>
WPI
Judgment Rule Efficacy
Accuracy from
. Weight on
Guilt by ; Guilt by
Association Association
ie
Actual Accuracy
(* :
Accuracy of Weight on
Guilt based on signs of
i rimina
Signs of Activity
Criminal Activity
Judgment Rule Efficacy
Sensitivity =
Guilty Investigated/
a Actual Guilty
Actual Accuracy /
Threshold or cutoff of
+ evidence for guilt
: Specificity = +
; Innocents Left Alone/
Actual Innocents
Testing the System
¢ Compare heavy to light
weight on Guilt by
Association
¢ Compare stingy congress to
expansive congress with
respect to resources
¢ Compare constraints on
protest and secrecy of
investigation to open society
Testing the System
¢ Measure sensitivity and
specificity
¢ Measure power of
investigative arm
¢ Measure resources required
to run investigations
Conclusions
¢ Pay serious attention to
sensitivity and specificity as
ee metrics of quality of
ee investigation
¢ Allocate resources liberally
¢ Allocate power conservatively
¢ Keep checks and balances in
place
Back to the Top \ { I |