Sample design requirements for the NCD Mobile Phone Survey have been developed so that unbiased population estimates of behavioral health risk can be generated for the target population of each country and for two analysis groups of interest defined separately by age or sex.
The allocation of the sample will be accomplished by explicitly stratifying the second-phase sample by age group (i.e., 18–29, 30–44, 45–59, and 60 years and older) and sex (i.e., male and female) with disproportionate allocation (to the mobile phone user population) among strata to produce a final respondent sample of mobile phone users that resemble the joint age group by sex distribution of the general population (as defined by the National Statistics Office or other reputable sources, such as the World Population Prospects: The 2012 Revision from United Nations Department of Economic and Social Affairs Population Division).
Designing the survey to generate precise joint estimates by sex and age group is preferred for the NCD Mobile Phone Survey, primarily to allow comparisons of estimates by these domains between different countries participating in this effort.
Deciding on a Sample Design
This manual provides a general statistical template for the randomly selected samples that should be used for the NCD Mobile Phone Survey. The purpose of this section is to present several options and design approaches currently relevant to the pilot testing phase of the survey protocol, where some countries have already done a NCD Mobile Phone Survey and others are looking to complete a NCD Mobile Phone Survey for the first time.
Sample Design Options for Countries Implementing the NCD Mobile Phone Survey Protocol
Three design options are considered in this setting:
Option 1. Two-phase sampling of MPNs from an implicit MPN frame.
- First phase – randomly sample MPNs from an implicit MPN frame of all possible MPNs.
Option 2. Two-phase sampling of MPNs from an explicit frame of user account MPNs.
- First phase – randomly sample MPNs from an explicit MPN frame of available MPNs provided by the telecommunication operators.
- Options 1 and 2. Second phase – sample stratified proportional to the general population but disproportionate to the mobile phone user population
Option 3. Stratified Multi-Stage Cluster Sample
- Split-sample household-level integration with the STEPS stratified multi-stage cluster sample
These sample design options are described in more detail in Section 3.