This document discusses how behavioral science and data science can be combined.
Behavioral science seeks to understand and change human behavior, while data science generally aims to understand and predict. Behavioral science can help data scientists look at data differently by considering behavioral biases, norms, and gaps between intentions and actions. It also suggests looking for different types of behavioral data. Finally, the document advocates explicitly using data analysis to design behavior change interventions and better understanding one's own cognitive biases.
14. 14
We overestimate our strength and ability,
each believing we’re “above average”
Which is reinforced with selective attention
and processing – confirmation bias
See Moore and Healy (2008)
Overconfidence Bias
16. An understanding of how people decide & act
What is behavioral science?
Tools to potentially help people do better
Rigorous testing to find results
16
17. 1. What is Behavioral Science?
2. How BeSci and Data Science Differ
3. How You Can Combine Them
4. Q&A
What’s Next
17
18. Data science generally seeks to understand
and to predict (and many other things)
The main difference:
Behavioral science seeks to
change human behavior
18
30. 30
To analyze descriptive norms,
you need a different a
different type of data
(reference groups, social
relationships, and level of
prior experience)
31. A Behavioral plan
Qual & quant research
that maps out the status
quo & the micro-
behaviors from there
to action.
To understand the
obstacles at each step.
31
32. 32
Condition Current Status? Obstacle?(Y orN)
Cue to think about taking
action
E.g. Relevance of email is unclear. E.g. Yes
Emotional
Reaction E.g. Aiming for positive. E.g. No
Conscious Evaluation of
costs and benefits
E.g. Long, multi-step sign-up process. E.g. Yes
Ability to act
(resources, logistics,
self-efficacy)
E.g. Ensure all users feel capable E.g. No
Timing and urgency
to act E.g. No urgency or commitment E.g. Yes
Prior Experience taking
action
E.g. Varied E.g. Not sure