Transcript Social norms and the promise of mass media entertainment

SOCIAL NORMS AND THE PROMISE OF
MASS MEDIA ENTERTAINMENT
Eliana La Ferrara
Invernizzi Chair in Development Economics
Bocconi University, Milan
May 16, 2016 - Mexico City
MOTIVATION
TV ownership (%)
Source: own calculations on DHS, 1987-2012
MOTIVATION
TV ownership (%)
Source: own calculations on DHS, 1987-2012
Women w/ secondary educ. (%)
QUESTIONS
Can we use TV & mass media for
development policy?
Is it effective?
 How does it work?
 What do we need to learn?

OUTLINE
1. Educational entertainment

Definition & theory
2. Evidence on effects

Information, preferences, social norms
3. Open issues

Content, social effects, timing
I. EDUCATIONAL ENTERTAINMENT
EDUTAINMENT
“Process of purposely designing and
implementing a media message to both
entertain and educate, in order to
increase audience members’ knowledge
about an issue, create favorable attitudes,
shift social norms, and change the
behavior of individuals and communities”
(Singhal and Rogers, 2004)
THEORY: ALBERT BANDURA

Social learning theory: individuals learn
through direct experience or experience
of others
 Media can exploit the reduced cost of
learning from others’ mistakes & allow
individuals to learn from models w/o
personal cost
THEORY: ALBERT BANDURA

Self-efficacy: belief in one’s capabilities to
organize and execute actions required to
manage prospective situations
 Media presents individuals w/
characteristics similar to target audience,
who succeed in reaching their objectives
 induce viewers to consider the goal
as feasible and increase motivation
THEORY: MIGUEL SABIDO ET AL.
Building blocks
 Differential role modeling: positive,
negative & transitional characters
 Identification: display situations/characters
similar to target audience (reduces counterarguing)
 Exemplification: show how to deal w/
problems or where to find information
CHANNELS OF INFLUENCE

Information: provide factual information
that is accessible & captivating


e.g., knowledge on contraceptives
Preferences & values: for given info, change
desirability of certain behaviors


e.g., already knew about family planning but
now wants fewer children b/c seen on TV
Shift mental models & perceived norms (WDR 2015)
II. EVIDENCE OF IMPACT
PREMISE ON METHODOLOGY

Early evidence mostly descriptive: opinions
of viewers who had chosen to watch


Cannot disentangle impact from “selection”
Empirically rigorous approaches have
recently relied on



Signal reception (topography, staggered entry)
Encouragement design to induce watching a
certain program on TV
Randomized screenings
II.A EDUTAINMENT WORKS FOR PROVIDING
INFORMATION
1. HIV PREVENTION IN NIGERIA
Banerjee, La Ferrara and Orozco (2016)

Shuga 3 (MTV) contains info on
 Risks of contagion from unprotected sex
& concurrent partners
 Testing: how it’s done; window period
 ARV drugs
RESULTS
 Sexual behavior
(-) concurrent partners
 (0) self-reported condom use
 (-) chlamydia among women


Testing
(+) knowledge of where to get tested
 (+) knowledge of window period
 (+) probability of testing


Treatment

(+) knowledge of ARVs
2. FINANCIAL EDUC. IN SOUTH AFRICA
Berg and Zia (forthcoming)


Making financial decisions can be
challenging for low SES individuals
Scandal! worked w/ National Debt
Mediation Association to insert a 2-month
subplot on debt & gambling
 Hire-purchase agreements
 How to ask for help from NDMA
RESULTS
+10% knowledge of hire-purchase
agreements, hidden interest, etc.
 (0) general financial literacy

-20% use of hire-purchase
 -17% gambling
 Shift in borrowing from informal to formal
sources


+15% intention to seek financial advice
from formal sources, but only in short run
(zero effect 4 months later)
3. PUBLIC WORKS IN INDIA
Ravallion, van de Walle, Dutta, Murgai (2015)
Low take-up of NREGS in Bihar
 25-min. movie w/ info on how the scheme
works:
 eligibility
 application process
 benefits

RESULTS
 Knowledge

(+) knowledge about program features, e.g.


Perceptions
(+) share who think NREGS has
 reduced migration (from 14% to 23%)
 improved work opportunities for their
family (from 12% to 22%)


# days (+50%); wage rate (+33%)
Usage

No impact on actual usage:

take-up rate, # days worked, wage
II.B EDUTAINMENT (PARTIALLY) WORKS FOR
CHANGING PREFERENCES AND NORMS
1. NOVELAS AND FERTILITY IN BRAZIL
La Ferrara, Chong, Duryea (2012)

During 1965-1985 Rede Globo’s novelas
portrayed women with few children &
desirable life-style  potential role models

After Globo’s signal reaches a municipality,
fertility rates decline


Effect stronger for women w/ low SES
Effect size is similar to giving women 1.6 more
yrs of education
2. CABLE TV & WOMEN’S
EMPOWERMENT IN INDIA
Jensen, Oster (2009)






Penetration of cable TV across Indian
villages in 2001-2003
Exposure to Western values
(-) acceptability of domestic violence
(-) son preference
(+) decision making autonomy
(+) girls’ education
BUT…
3. SHUGA IN NIGERIA
Banerjee, La Ferrara and Orozco (2016)

Mixed evidence on attitudes affected
 (+) attitudes towards HIV+ people
 No generalized improvements in
acceptability of domestic violence
(effects on selected attitudes)
4. RECONCILIATION IN RWANDA
Paluck (2009)

Radio soap opera designed to address
mistrust, lack of communication, trauma
 No impact on beliefs re: sources of
violence
 (+) impact on prescriptive norms
(how people should behave)
 No impact on descriptive norms
(how people actually behave)
SUMMING UP WHAT WE KNOW SO FAR

Evidence that EE programs are effective
in conveying information

Mixed results when it comes to
preferences, social norms & behavior

… difficult to inderstand why from
existing evaluations
III. WHAT WE NEED TO LEARN
III.A WHAT CONTENT TO DELIVER
AND HOW?
INFORMATION VS PREFERENCES

If main channel through which EE works is
information, then need to fill content w/
factual details

If main channel is preference change,
may want to invest more in creating
emotional connection/empathy

… things can go wrong if you omit or
emphasize the wrong things
EXAMPLES OF UNINTENDED EFFECTS

NREGS movie in India
Selective information: main character was
a man
 (-) perceived eligibility of women among
people who watched movie!

EXAMPLES OF UNINTENDED EFFECTS

Entrepreneurship in Tanzania
(Bjorvatn et al., 2015)
EE show on lives of 6 young entrepreneurs
 (+) aspirations to become entrepreneurs
 (+) entrepreneurial traits & probability of
starting a business
but
 (-) education: higher failure rate in exam
& (-) continuation after high school
 show emphasized work & not education

OTHER UNINTENDED EFFECTS

Consumer demand & mix of educational
& non-educational content:

How much intrigue, gossip, etc. are we
willing to tolerate to pass on an educational
message?
ETHICS
Border b/w EE & instrumental
manipulation?
 How to evaluate welfare when you are
changing people’s preferences?


Which values is it fair to affect?
III.B SOCIAL NORMS & THE ROLE OF PEERS
INTERACTION & COORDINATION

Settings where my optimal action
depends on what others will do, e.g.,
Pay bribes or not?
 Ask partner to wear a condom?


Not enough to show the viewer what
he/she should do  need to change
their beliefs about what others will do
EXAMPLES

Some of the previous failures to change
norms may be due to failures to shift
beliefs on what other people think/do,
e.g.,
Trust in Rwanda
 NREGS take-up in India
 Domestic violence in Nigeria

UNDERSTANDING SOCIAL EFFECTS
“Announcement” design in our Shuga
evaluation
 Collect info on baseline beliefs about
others’ behavior, e.g.,


How may people in your community would
tell their partner if they were HIV+ ?
After Shuga screening, show them a
video w/ interviews & statistical info
collected from other respondents
SUPPOSE YOU HAD HIV. WOULD YOU TELL YOUR
PARTNER?
Effect of
announcement should
be positive for
respondents whose
prior was < 66%
PRELIMINARY RESULTS
We find “announcement effects”
consistent w/ our hypothesis, mostly in
the short run
 Effects don’t differ if respondent was
allowed to bring a friend to the
screening
  How to credibly induce collective
shift in beliefs to change social norms?

OTHER UNANSWERED QUESTIONS

How large are spillovers, i.e., effects on
people who did not watch movie?

Interplay of “traditional” and “new”
norms: how to make social change
acceptable?
III.C TIME
DURATION & FREQUENCY OF EE
PROGRAMS

Experimental evaluations (RCTs) of EE
programs involve a single movie or a
limited no. of episodes

How long should EE run for?

At what frequency?


Info may be forgotten if freq. too low
But it takes time to learn & change preferences
HOW PERSISTENT ARE THE EFFECTS?

Most experimental evaluations estimate
short run (or medium run) effects

Long run evidence mostly comes from
non-experimental studies (exception:
Bjorvatn et al.)

Investing in EE more worthwhile if effects
are long lasting  need long term
follow-up of experimental studies
IV. CONCLUSIONS
TAKEHOME MESSAGES
1. EE works for providing information.
Evidence on changes in norms & values
less clear cut
2. Need rigorous evidence on how to
design content to maximize impact &
minimize unintended effects
TAKEHOME MESSAGES
3. Need to understand role of coordination
& peer effects to target individual or
collective beliefs
4. Need to better understand the time
dimension:
- intensity/duration of exposure to EE
- short run vs long run impacts
If you haven’t had enough…
E. La Ferrara, “Mass Media and Social Change:
Can we Use Television to Fight Poverty?”, Journal
of the European Economic Association, forthcoming
Available at:
www.igier.unibocconi.it/laferrara
THANK YOU!