GWI Core has three components; the main survey, the mobile survey, and the brand and media module.
Most respondents complete the main survey, but a significant proportion complete the mobile survey instead. The mobile survey is run in parallel to the main survey, but contains slightly fewer questions and has been optimized for smaller screens.
Approximately half of the main survey respondents also complete the brand & media module - a recontact survey containing additional questions on brand engagement and media consumption.
After fieldwork, these three surveys are combined and presented as a single data set in our platform.
Why do we run a separate mobile survey?
The growing availability of smartphones has boosted internet penetration the world over, particularly in emerging markets, where most internet users are mobile-first and where significant minorities might be mobile-only. The mobile survey was launched in 2017 to help us capture these users and contains a selection of questions from the main survey in a format optimized for mobile. In Ghana, Kenya, Morocco and Nigeria, we only run the mobile survey.
Why do we have the brand & media module?
By asking about brand engagement and media consumption in a separate survey, we're able to cover a wider variety of subjects and go into greater depth than would possible in a single survey. This means we're able to collect a large number of data points on our respondents without exhausting them and reducing the quality of the answers we receive.
What does this mean for users?
As all questions from the mobile survey are also asked in the main survey, and because all respondents who’ve completed brand and media module have also completed the main survey, all questions can be used with each other, regardless of which survey(s) they're asked in.
Additionally, responses to all questions are weighted to represent the same overall number of internet users aged 16-64 in each market. This means that responses to some questions will have a greater weight applied to them than others. For instance, because only half of the main survey respondents complete the brand and media module, responses from this section will have a larger average weight.
This can sometimes mean that a data point from the brand and media module has a larger universe figure but a smaller sample size than a data point from the main survey. This isn't something to be concerned about, as both surveys have large and robust samples in each market. Additionally, the calculations that make this work are completed seamlessly in the platform, meaning no action is required.
Can I apply a base of brand and media module respondents only, should I need to?
Yes - by creating an audience and selecting all response options to a brand engagement question (for example, by creating an audience of people who currently have an Audi, have had an Audi previously or have never had an Audi) and then applying it as a base. This would allow you to look at a question from the main survey whilst excluding all respondents who didn't complete the brand and media module. However, generally, we'd advise against doing this, as our platform will always use the largest available sample for each data point to ensure the displayed figures are as accurate as possible.