Inducing Complete Coverage for Immunizations by Channeling Social Media

Last registered on January 02, 2018

Pre-Trial

Trial Information

General Information

Title
Inducing Complete Coverage for Immunizations by Channeling Social Media
RCT ID
AEARCTR-0000757
Initial registration date
January 02, 2018

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
January 02, 2018, 4:28 PM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Primary Investigator

Affiliation
MIT

Other Primary Investigator(s)

PI Affiliation
Stanford University
PI Affiliation
World Bank
PI Affiliation
Microsoft Research
PI Affiliation
World Bank

Additional Trial Information

Status
On going
Start date
2015-05-22
End date
2018-12-31
Secondary IDs
Abstract
Indonesia’s performance on the health Millennium Development Goals (MDGs) still remains poor. On MDG 4, which explores the mortality rate of children 0-5, Indonesia has made some progress, moving from 97 per 100,000 live births in 1990 to 44 per 100,000 by 2010. However, the rate of progress has declined, and it remains unclear whether the MDG will be met.

One of the key drivers of the progress for child mortality is the rate of complete immunization of children under 5 against a range of childhood diseases, including measles, diphtheria, whooping cough (pertussis), tetanus, tuberculosis, polio and Hepatitis B. Despite their importance, complete immunization coverage against these diseases is quite low in Indonesia – for instance, only 27% of children are fully immunized against polio, and more than 21% of children have not been fully immunized against measles.

The Government of Indonesia has made significant efforts to increase the supply of health services to improve these outcomes. By 2011, only 6% of rural children lacked access to a primary health care center, and almost no urban children lacked access. Only 8% of rural children were born in a village without a midwife, compared with only 2% of urban children. Despite important progress on the supply side, low rates of care-seeking persist and outcomes remain poor.

The Government is seeking to investigate whether electronic media -- social media such as Twitter, as well as conventional text messaging to cell phones -- can help increase vaccinations. In Indonesia, use of Twitter and other social applications – especially those accessible via mobile – has skyrocketed. Indonesia holds the world’s fifth largest number of Twitter accounts, though recent growth has outpaced three of the four nations with larger user bases: Brazil, Japan, and the United Kingdom. On a daily basis, we observe between 400 and 500 million tweets, of which about 75% are original tweets and the remainders are retweets. The Government thus believes that Twitter has an important potential role in spreading social information, such as immunization campaigns.

In addition to the policy interest, there is also a separate set of academic questions about the diffusion of information over networks: how we communicate, on what topics, and whether this induces actions. With the proliferation of new social tools, individuals engage in information exchange in more ways than ever before, and this has implications on political beliefs, take up of social policies, preferences related to care seeking behaviors, among other areas.

This research project will therefore investigate whether disseminating social messages on Twitter and direct text messages (SMS) can have a positive and significant impact on take-up of socially positive behaviors, such as immunization seeking behaviors.
External Link(s)

Registration Citation

Citation
Alatas, Vivi et al. 2018. "Inducing Complete Coverage for Immunizations by Channeling Social Media." AEA RCT Registry. January 02. https://doi.org/10.1257/rct.757-1.0
Former Citation
Alatas, Vivi et al. 2018. "Inducing Complete Coverage for Immunizations by Channeling Social Media." AEA RCT Registry. January 02. https://www.socialscienceregistry.org/trials/757/history/24657
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
The interventions are social messages related with immunization delivered over Twitter by celebrity accounts.
Intervention (Hidden)
I. Twitter
We will recruit approximately 40 celebrities, several (up to 20) known organization related with immunization (e.g., Ministry of Health, Office of President's Special Envoy for MDGs), and up to 1,000 ordinary individuals who are Twitter users who are not celebrities.

The design cross-cut the following treatments:
Treatment 1: Who originates the tweet?
Tweets can either come from (i) celebrity directly tweeting, (ii) celebrity retweeting a tweet from organizations, (iii) celebrity retweeting a tweet from ordinary users, or (iv) ordinary users retweeting a tweet from another ordinary user previously retweeted by a celebrity.
Treatment 2: Is source given for information?
We compare tweets with info from celeb alone, vs. info with a source. E.g., “World Health Organization” or “UNICEF” or URL included in the message.
Treatment 3: What is tweeted?
Tweets are randomized from 1000+ possible tweets with the following content types: fact, importance, or access message. All contents are hashtagged #AyoImunisasi.
Treatment 4: When are the tweets sent out?
Every celebrity is randomly assigned to tweet anywhere from 0 to several times a day. Multiple tweets can also be sent simultaneously in the setup.


On each day of the experiment, every celebrity will be randomly assigned to a cell. We will run up to 100 days of the experiment. An assignment to a cell consists of a triple of: (i) who tweets, (ii) what is the schedule, (iii) whether the tweet has a credibility boost.
Intervention Start Date
2015-07-02
Intervention End Date
2016-01-21

Primary Outcomes

Primary Outcomes (end points)
(a) propensity to retweet information
(b) knowledge and discussion about immunisation
(c) recent immunization behavior
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
This research project investigates whether disseminating social messages on Twitter can have a positive and significant impact on take-up of socially positive behaviors, such as immunization seeking behaviors. The project involves a Twitter campaign.

Approximately 300 users (including celebrities, immunization-related organizations and non-celebrity Twitter users) are recruited for the Twitter campaign. All of these twitter accounts will publish tweets on immunization, but the tweets will be varied based on (1) whether a celebrity, organization, or ordinary individual is the first to tweet the message, (2) how often the tweet is scheduled to be published, and (3) whether the tweet includes a credibility boost (in the form of mentioning an official immunization-related organization)/content of the tweet.

After the treatments have been implemented, data will be collected from approximately 2000 households via phone surveys, and researchers will also observe the spread of the Twitter campaign within its users.
Experimental Design Details
Randomization Method
Randomization will be done using Matlab
Randomization Unit
The celebrities are randomly split into two phases of campaign. Additionally, we also randomly assign the type of treatments for each tweet sent in the campaign.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
N/A
Sample size: planned number of observations
approximately 40 celebrities and all their followers (approx. 5 million)
Sample size (or number of clusters) by treatment arms
Approximately 40 celebrities in the total sample, split in two phases of campaign.
Approximate number of tweets in each phase: 450 tweets
Approximate number of tweets by tweet source: 400 direct tweets, 250 joe tweets, 250 org tweets
Approximate number of tweets for each credibility boost type: 300 tweets
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
N/A
IRB

Institutional Review Boards (IRBs)

IRB Name
COUHES MIT
IRB Approval Date
2014-06-19
IRB Approval Number
1406006433
IRB Name
Stanford University
IRB Approval Date
2014-06-04
IRB Approval Number
31451

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials