Our study sample consists of 2,000 smokers aged 18 or older who reside on the island of Mindanao in southern Philippines. Green Bank marketers identified smokers by approaching people and asking them whether they smoke regularly. If they did, the marketer then asked if they wanted to participate in a short survey on smoking. All subjects received an informational pamphlet on the dangers of smoking, and a tip sheet on how to quit. Since the primary objectives of this study were to determine whether first there was demand for CARES, and second wh ether CARES increased smoking cessation, the marketers only collected very quick and basic baseline data on age and smoking status (see Section V-A for more details).
The experiment was implemented in three distinct waves of marketing. The first two waves took place in Butuan City from August to December 2006. After completing the baseline survey marketers revealed a sticker on the back of the survey that randomly assigned the subject to one of four groups: (1A) CARES with deposit collection, (1B) CARES without deposit collection, (2) Cues, or (3) Control. 15 The probability of assignment to groups was initially 45%, 45%, 5%, and 5%. After establishing that there was sufficient takeup of CARES, Green Bank changed the assignment probabilities to 15%, 15%, 30%, and 40% for the second wave. 418 smokers were surveyed (and hence drawn into the sample frame) in the first two waves. Of the 266 assigned a CARES offer, 34 took the product. Two individuals from th e Cues group also opened an account (after hearing about the product and approaching bank staff). In our analysis we code these individuals in the Cues group, in adherence to the random assignment.
|Source Institution||World Bank|
|Place of Publication||Pasig City|
|Original Publication Date||December 1, 2008|