Satish Nargundkar Georgia State University Framing and Cognitive Bias Positive Framing Risk Averse behavior Negative Framing Risk Seeking behavior Tversky Kahnemann 1981 Framing ID: 708694
Download Presentation The PPT/PDF document "Framing the Analytics Problem" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Framing the Analytics Problem
Satish
Nargundkar
Georgia State UniversitySlide2
Framing and Cognitive Bias
Positive Framing
Risk Averse behaviorNegative Framing Risk Seeking behaviorTversky, Kahnemann (1981)
Framing
Treatment A
Treatment B
Positive
"Saves 200 lives"
"A 33% chance of saving all 600 people, 66% possibility of saving no one."
Negative
"400 people will die"
"A 33% chance that no people will die, 66% probability that all 600 will die."Slide3
Framing as an Analytics Problem
Identifying the Business Objective
Converting into a quantitative problem
Key Tasks:Identifying a dependent (outcome, performance) variableIdentifying an outcome periodCreating a sampling planSlide4
Business Objectives - Prioritization
Warmup Exercise: Quality of Marker Pen
What does it mean?
How would you quantify?How would you prioritize?Use Mini-casesSlide5
Planning to get the right dataSlide6
The Dependent Variable
Example 1: Education
Predict the likelihood that a high school graduate will enter a 4-year degree program within 3 years of graduating from high school.
What is the dependent variable?What values will it take on?What is the outcome period?What data would you collect? From When? Will the time frame be the same for the Y and the Xs?Slide7
Answer to Example 1
Status of HS graduate 3 years after graduation
Values: 1 = Entered College; 0 = Did not enter College
Outcome Period: 3 yearsSample Data
2013
Outcome Period
2016
2008Slide8
Example 2: Financial Services
Build a model to predict customer risk. If we were to accept a person as a customer (provide a loan
or credit card, for instance) how likely is the customer to default within the next 12 months? Key points: Status vs. RiskHow far back?Exclusions?Slide9
Example 3: Criminal Justice
Predict
the likelihood
that a person released from prison after serving at least a 5-year sentence for being convicted of a violent crime is likely to commit another such crime within one year of being released. A released criminal may commit a violent crime again, or commit a
non-violent crime
, or
stay clean
(
no crime
of any kind) in that year after release. You believe that since some of
the laws
regarding violent crimes changed significantly in the year 1996, people
convicted anytime
during or before then are not relevant to the study.Slide10
Conclusion
Students generally need
P
ractice identifying key variables Deciding on the sample needed Emphasis on the time frameSlide11
Questions?
Thank you!