Promoting Venezuelan migrants' registration to Sisbén

The study has two main phases. In the first, we seek to evaluate the effect of different information nudges on Sisbén registration. In the second phase, we aim to identify the dynamics of participants’ social capital and if it correlates with Sisbén take-up or not.

The first stage consists of an RCT. We will send the participants a baseline survey that includes screening questions. We will assign eligible participants, who from now on we will call “Generation 0”, into three groups, one control and two treatments (see Figure 1). Participants assigned to the control group will not receive any information, besides the baseline and endline survey. Participants in Treatment 1 group will receive informational nudges via SMS. Participants in Treatment group 2 will receive information through WhatsApp. The content of the messages is the same for the two treatment groups.

The second phase of the study will complement the first phase’s intervention by mapping out migrant networks and information diffusion among those networks. In this second phase, we will send eligible participants -participants that already requested Sibén- three to four messages inviting them to ask other Venezuelans to contact us and be part of the study. From now on, we will refer to this first group of participants as “Generation 1.” Each participant from Generation 1 will receive a unique code that referrals must send us through WhatsApp to register in the study.

2022-10-11

2022-11-30

The primary outcome variable is a binary indicator that is equal to 1 if the individual has registered in Sisbén and zero otherwise.

The purpose of the intervention is to encourage unregistered Venezuelan migrants to register in Sisbén through information provision. Given the eligibility criteria of the intervention, none of the selected participants should be registered at the beginning of the trial. The main outcome is therefore whether the intervention motivates migrants' registration into the Sisbén database.

We also have secondary outcomes that are "quiz" questions to measure disinformation with respect to Sisbén. These are binary indicators on several myths that are 1 if the individual responds correctly and zero otherwise.

In qualitative work, several myths about Sisbén were commonly found, specifically regarding its nature (conceptions of it not being an information system, but a social program or other aid) and the required documents to request registration. We will ask participants their perceptions through quiz questions to understand whether the provided information helps eliminate these myths.

For the first phase of the study, given the sample and the goals of the intervention, we randomly assign treatment at the individual-level, allocating persons into one of three groups: i) control group, which does not receive information on Sisbén registration; ii) a group that receives information on the Sisbén registration process via SMS; iii) a group that receives the same information on Sisbén registration via WhatsApp.

Given eligibility criteria and consent, 860 individuals will be part of the RCT. We assign one third into each of the three groups, with sample sizes of 303 (Control), 280 (Treatment 1), and 277 (Treatment 2).

Randomization done in office by computer (Code available upon request)

Individual

No

860 individuals

860 individuals

We assigned one third of the planned number of observations into each of the three groups, with final sample sizes of:
303 (Control)
280 (Treatment 1)
277 (Treatment 2)

We conduct power calculations by simulation since we have no pre-intervention data on Sisben registration for these eligible individuals (it is zero for all). We then split these into the three groups and simulate a binary outcome with a normal distribution with mean 0.10 and standard deviation 0.4, calculating power for different effect sizes (from 1 to 20 percentage points). The minimum detectable effect with our sample size is 0.048, so we can detect a change of 4.8 percentage points or higher in Sisbén registration. For the secondary outcomes relating to quiz questions asked in a survey, we may detect a change of 10 percentage points on binary variables that identify correct answers.

Pre-Analysis Plan