|
Field
Trial Start Date
|
Before
February 04, 2018
|
After
September 05, 2018
|
|
Field
Trial End Date
|
Before
December 14, 2018
|
After
March 14, 2019
|
|
Field
Last Published
|
Before
June 17, 2018 02:56 AM
|
After
October 03, 2018 04:56 PM
|
|
Field
Intervention (Public)
|
Before
There are two interventions that we will implement for two health behaviors in various combinations across treatment arms. The first is a messaging program: subjects will receive two daily messages about the behavior, one that is a simple reminder, and another that contains information about the behavior's benefits. The second is an incentive program: participants will earn one lottery ticket for every day they successfully do the behavior, and winning tickets will be drawn at the end of the treatment period. Both interventions will last 4 weeks.
There will be two additional interventions that we will use to identify mechanisms. First, some participants will receive a messaging program with fun health facts that correspond to no action. Second, for some participants, the framing of treatment assignment will make salient the fact that we believe this assignment will be beneficial for them (as opposed to making salient the random nature of assignment).
|
After
There are two interventions that we will implement for two health behaviors in various combinations across treatment arms. The first is a messaging program: subjects will receive two daily messages about the behavior, one that is a simple reminder, and another that contains information about the behavior's benefits. The second is an incentive program: participants will earn one lottery ticket for every day they successfully do the behavior, and winning tickets will be drawn at the end of the treatment period. Both interventions will last 4 weeks.
|
|
Field
Intervention Start Date
|
Before
June 18, 2018
|
After
September 06, 2018
|
|
Field
Intervention End Date
|
Before
August 31, 2018
|
After
November 30, 2018
|
|
Field
Experimental Design (Public)
|
Before
Some participants will receive a messaging programs for only one behavior, some will receive messaging programs for both behaviors, and some will receive the messaging program for one behavior and the incentive program for one behavior. Some participants will receive the message program for one behavior plus the "distractor" message program that does not correspond to any behavior. Finally, for some participants, we will make salient the fact that we believe this particular treatment assignment will be beneficial for them. The key outcome of interest will be whether or not participants engage in each action at the individual-day level. This experiment design will allow us to identify various mechanisms of interference between behaviors and interventions.
|
After
Some participants will receive message or incentive programs for only one behavior, some will receive messaging programs for both behaviors, and some will receive the message program for one behavior plus "distracting" messages that do not correspond to any behavior. The key outcome of interest will be whether or not participants engage in each action at the individual-day level. By looking at spillovers and interactions between interventions, we will be able to distinguish between two types of limited attention -- internal and external -- with important policy implications.
|
|
Field
Randomization Method
|
Before
In office by a computer
|
After
Randomization is done within the Qualtrics survey, stratified on six variables (see Pre Analysis Plan)
|
|
Field
Planned Number of Observations
|
Before
1920
|
After
3000
|
|
Field
Sample size (or number of clusters) by treatment arms
|
Before
240 individuals will be assigned to each of the 8 treatment groups
|
After
Each treatment arm will have a slightly different size, computed in a power calculation that depended on various factors (minimum detectable effect, serial correlation of the outcome, variance of the outcome).
|
|
Field
Power calculation: Minimum Detectable Effect Size for Main Outcomes
|
Before
The minimum detectable effect size for crowd-out is 5.8 percentage points (the effect size I found in a pilot), or 0.16 standard deviations. This constitutes a 37% reduction in the efficacy of a program (from raising the probability of action by 15.8 percentage points, to raising the probability of action by only 10 percentage points).
|
After
The minimum detectable effect size for both spillovers and crowd-out is 6 percentage points, or about 0.18 standard deviations. This constitutes a 50% reduction in a behavior, relative to the control (in the case of spillovers), or a 30% reduction in the efficacy of the program (in the case of crowd-out).
|
|
Field
Secondary Outcomes (End Points)
|
Before
Score on quiz about information sent via messages
|
After
Opt-outs, score on quiz about information sent via messages
|
|
Field
Secondary Outcomes (Explanation)
|
Before
The score on the quiz will be the fraction of questions answered correctly in the endline survey. We will then construct an index that is normalized around the control group.
|
After
Opt-outs are whether or not participants opted out of the SMS programs at any point. The score on the quiz will be the fraction of questions answered correctly in the endline survey.
|