Live follow-up game? - A Path That`s Clear

Using Giving Games to Study and
Improve Charitable Giving
Brown Bag Presentation at ICES by
Jon Behar of APathThatsClear.com
April 4, 2013
Introduction
• Agenda
• My Background
• Questions: Why don’t people donate better?
Is there a program that could get them to?
The basics of Giving Games
• Donation decision is delegated to players
• Players select from a restricted set of charities
• Opportunity set is structured to provoke
thought
• Designed with an intent to improve giving
behavior
The potential benefits from better
giving
• $300b annual US donations, $200b from
individuals
• Aggregate impact a function of both dollars
and “good” per dollar
• Huge variations in impact across charities
equates to enormous opportunity
Why don’t donors “give well”?
•Scope Insensitivity: Donors don’t respond to the
scale of the problem
•Identifiable victim effect:
Tangible victims are more evocative
than statistical victims.
•Psychic numbing: thinking about suffering, particularly
large scale suffering, desensitizes people.
•Immediacy biases: recently received information carries
more emotional weight
Making people aware of these biases
can crowd out generosity
Cryder: “It seems almost as if any method of priming
a deliberative mindset… leads to less generosity.”
How donors allocations deviate from
utilitarian standards
• The waste heuristic: preoccupation on
“efficiency” rather than impact
• Focus on Average, not Marginal, impact
• Diversification heuristic: give to many orgs
instead of just the best
• Parochialism: favoring in-groups (e.g. conationals)
• Tax vs. Charity framing: preference for
voluntary giving
What actually motivates donors?
Segment
How they give
% Population
% Donations
Repayer
I support organizations that have
had an impact on me or a loved
one.
23%
17%
18%
18%
High Impact
I primarily give to well known nonprofits through a payroll deduction
at work.
I give to the nonprofits I feel are
generating the greatest social
good.
16%
12%
Faith Based
We only give to organizations that
fit with our religious beliefs.
16%
18%
See the Difference
I think it's important to support
local charities.
14%
10%
Personal Ties
A lot of my giving is in response to
friends who ask me to support
their causes
13%
25%
Casual Giver
How donors “research” charities
How to teach giving
• Making donors mindful of their own giving
criteria makes them use those criteria more
• Experiential philanthropy courses
– Learning by Giving Foundation ($10k/class)
– Once Upon a Time Foundation ($100k/class)
The Cast of Characters
• Funder: “an anonymous sponsor”
• Players: College students (future donors) with
group identity
• Organizers: “Effective Altruism” chapters,
experiential philanthropy students, faculty
Breakdown of Giving Games
Type
# of Games Total # of players
Group
Discussion/Activity
11
129
Total $
2025
Giving Stall
4
907
975
Online
3
78
900
Total
18
1114
3900
Results: Group discussions
Description
Organizer
#
Players
Vanderbilt RA
VU Undergrad
17
Safe Haven (10) vs. AMF (7)
$250
Intern Program
VU Undergrad
14
$250
A Capella (with friend)
GWWC: Princeton
11
A Capella (no friend)
GWWC: Princeton
7
Freshman Floor
GWWC: Princeton
8
FSD (14) vs. AMF (0)
AMF (8) vs. Group regranting (2) vs.
Group's Fund (1)
AMF (6) vs. Group's Funds (1) vs.
GiveDirectly (0)
AMF (7) vs. local KIPP (1) vs. Group's
social budget (0)
Charities (Votes)
Prize
$250
$100
$100
Survey results from Group Discussions
Changed your thinking?
Game Successful?
17
32%
88%
Intern Program
14
7%
43%
A Capella (with friend)
11
55%
77%
A Capella (no friend)
7
86%
93%
Freshman Floor
8
88%
100%
Avg. across players:
45%
77%
Avg. across games:
53%
80%
Description
# Players
Vanderbilt RA
Survey results: Follow-up activities
Description
Brief
Receive
Share
online Live followGiveWell GiveWell via
follow-up up game?
research?
FB?
game?
Attend
meeting?
Accept any Organize Giving
followup?
Game?
Vanderbilt RA
#N/A
#N/A
#N/A
#N/A
#N/A
#N/A
62%
Intern Program
A Capella (with
friend)
A Capella (no
friend)
#N/A
#N/A
#N/A
#N/A
#N/A
#N/A
46%
73%
55%
45%
27%
73%
91%
#N/A
57%
0%
43%
0%
14%
71%
#N/A
Freshman Floor
100%
88%
63%
25%
38%
100%
#N/A
Avg. of Players
77%
50%
50%
19%
46%
88%
55%
Avg. of Games
77%
47%
50%
17%
42%
87%
54%
Extrapolating to other Group
Discussion GGs
• Organizer feedback suggests other games had
impact similar to Princeton GGs
• Vanderbilt games likely understate expected
impact
• Princeton campus has atypical exposure to
effective giving message
Results: Giving Stalls
Location
Players
Charities (Votes)
Prize
Brown Activities Fair
21
AMF (17) vs. GiveDirectly (4)
$1/Person
34
Fred Hollows (15) vs AMF (12) vs
Make-A-Wish (7)
$1/Person
Harvard Activities Fair
UK Freshers Fairs (3-4
campuses)
Higher Education in the
21st Century Conference
~800
52 (~110
attendees)
GWWC: Cambridge’s Giving Stall
AMF (~640) vs SCI? (~160)
SCI (38) vs Wikipedia (14)
£200
£2/Person + £250
Bonus (50
participants)
Results: Online Games
• Selection bias makes interpretation difficult
• Consistent with Giving Stalls, players seem
willing to submit comments
• Some players will share games over social
media
• Transition to Facebook platform in the works
Observations from cross-game analysis
•
•
•
•
•
•
•
•
It’s plausible that GGs teach many players give better
GGs can have a significant impact on behavior
Players don’t see the money as theirs
Money matters, but mostly in a threshold sense
Organizer capability matters
Social ties between players facilitate good discussions
With Win/Win games, you can’t lose
GGs provide a great window into how donors think
GGs “jump to the solution”
Next Steps: More donors running
games
• Sources of funding
– My giving
– Individual donors
– Foundations
– Organizers
• Advantages
– Free to run
– Leveraged impact
– Emotional leverage
Next Steps: Researchers using GGs
• GGs add an intention to teach better giving to
existing research frameworks
• Field experiments through collaboration with
A Path That’s Clear or other funders
• Giving Stalls” or Online GGs most promising models
• Data sharing will facilitate meta-analysis
Key questions to pursue
• What long-term impact do GGs have?
• What are the best ways to mitigate the impact of
“bad heuristics” and propagate “good heuristics”?
• How sensitive are outcomes to the artificial
constraints of the GG model?
• How do the process inputs (players, prize,
charities, activity) translate to the process
outputs (votes, discussion, follow-ups)?
• What are the key drivers of “viral variables”?
• What strategies should a GG sponsor use?
In conclusion…
• Let’s collaborate on a field experiment!
• Look for opportunities to research using GGs
• If your network includes anyone who’d be
interested in GGs, please put us in touch
• If you give, please give well
• Any and all feedback is welcome