NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Impacts of Access to the Internet
Last registered on October 29, 2018


Trial Information
General Information
Impacts of Access to the Internet
Initial registration date
October 28, 2018
Last updated
October 29, 2018 5:18 PM EDT
Primary Investigator
MIT Sloan School
Other Primary Investigator(s)
Additional Trial Information
On going
Start date
End date
Secondary IDs
According to recent studies, the mobile internet in Kenya is among the fastest globally. There is an assumption that high-speed Internet has an impact on people’s lives as new Internet users have access to a vast amount of resources and can improve their own skills, employability and access to jobs. However, there is no research that illustrates the impacts of mobile Internet access. The purpose of the study is to understand how access to the Internet affects people’s lives. This project will study how access to data-enabled phones provide access to high speed mobile internet for people and how this affects peoples’ social and economic lives, whether they are able to search for job opportunities, learn skills, etc. over the internet.
External Link(s)
Registration Citation
Suri, Tavneet. 2018. "Impacts of Access to the Internet." AEA RCT Registry. October 29. https://doi.org/10.1257/rct.2550-1.0.
Former Citation
Suri, Tavneet. 2018. "Impacts of Access to the Internet." AEA RCT Registry. October 29. http://www.socialscienceregistry.org/trials/2550/history/36426.
Sponsors & Partners

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

Request Information
Experimental Details
Safaricom implemented a randomized controlled trial based on administrative data that provides individuals with access to free (internet) data. They created two treatment groups and a control, with 2,000 individuals in each group. The two treatment groups were:

1: 5GB of free data, valid for 90 days and renewed 4 times
2: 15GB of free data, valid for 90 days and renewed 4 times

Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
Use of the internet and social media, social networks, attitudes (social and political), aspirations and economic opportunities (employment, migration)
Primary Outcomes (explanation)
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
There are two treatment groups that get free data (these are individuals who have never used data before). One group gets 5GB for free and the second gets 15GB for free. The free data offers were valid for 90 days. The offers were then repeated four times so that the treated individuals had free data for a year.

The sample from which the treatment and control groups were chosen was restricted by Safaricom as follows:

(i) Safaricom customers who have data enabled phones but who have not used data before (those with 0MBs of data use in the 18 months prior to the experiment and those not in the top percentile of data expenditures (a small number of individuals had purchased data at some point, even though they had not used any of it).

(ii) Among this sample, they then dropped those with below median revenues in the last 2 months, or with below median revenues in the last 6 months or those with less than median voice revenue in the last 2 months. The aim of this restriction was to ensure that these customers were active users of Safaricom and not SIM cards that are out of service or not used much.

From this final sample of customers, Safaricom chose a random 4,000 to receive free data. Half of these 4,000 were in the 5GB treatment group, and half in the 15GB treatment group.

These 4,000 will be compared to a control group of 2,000 chosen also at random from the same sample of customers described above. Should we need a bigger control group for power, we can go back to the original sample and add more consumers to the control group as the treatment individuals were chosen at random from the described sampling frame.
Experimental Design Details
Randomization Method
Randomization done in an office by a computer (using Stata code)
Randomization Unit
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
4,000 individuals treated
Sample size: planned number of observations
6,000 individuals total (though the control group can be expanded as the study group was chosen at random from a sub-population of customers)
Sample size (or number of clusters) by treatment arms
2,000 individuals get 5GB of free data, valid for 90 days and renewed 4 times
2,000 individuals get 15GB of free data, valid for 90 days and renewed 4 times
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB Name
MIT Committee on the Use of Humans as Experimental Subjects
IRB Approval Date
IRB Approval Number
Post Trial Information
Study Withdrawal
Is the intervention completed?
Is data collection complete?
Data Publication
Data Publication
Is public data available?
Program Files
Program Files
Reports, Papers & Other Materials
Relevant Paper(s)