Pilot SMS Survey Reaches 36,000
Monday, 22 August 2011 19:23
Print
ICTs

As part of a Major research project investigating the local effectiveness of the HYOGO framework, Gamos has analysed data gathered through innovative SMS based surveys. 36,000 reposes were gathered from amongst those living in disaster affected countries. In partnership with GNDR, Gamos has analysed the experience of this eSurvey project to identify the critical success factors. Along with many other benefits, this survey technology has the potential to maximise the opportunity created by dramatic recent increases in mobile phone usage in many low to middle income countries.

Organisations partnering in this particular piece of research estimate the cost of face to face surveys at $125/survey (this includes training, mobilising communities, transport and logistics). The mobile phone based eSurvey has the potential to lower the cost of conducting a simple survey by an order of magnitude – the cost in this instance worked out at just under $5/survey.

Observations from the pilot project include:

  • Negatively, an eSurvey questionnaire does not generate a sense of corporate identity in the way interviews can. During the VFL process, face to face interviews turned out to be a key activity in stimulating dialogue and mobilising communities to push for change at a local level.
  • The first phase of the eSurvey tended to reach young, urban males reflecting patterns of mobile phone ownership (subscriber rates among women in low and middle income countries are 79% of subscriber rates among men) and internet access (in Africa, twice as many men use the internet as women.
  • Airtime credit plays a big part in the determining who completes the survey. For example, in Kenya 40% of mobile users don’t keep a balance on their mobile phone. This means creative means and/or agreements with networks need to be used to reach the more marginalised sectors of society. 99% of completed questionnaires were done so via the project webpage. Only in Kenya did participants complete the survey using only their mobiles. Kenya was the only country where an agreement was reached with the networks to make the completion of the questionnaire free to the respondent. This highlights the importance of the aforementioned airtime credit.
  • English is being adopted as the language of texting in several countries. (For example in Bangladesh people texted in a basic English vocabulary in preference to Bengali, which required multiple keystrokes). This does not mean that texting can overcome literacy barriers, but it does indicate that cross cultural research using mobiles may become easier than expected.
  • Further sampling issues included varied levels of network coverage and the self-selection of respondents. However, indicators to identify target respondents and selection rules can be developed at the survey design stage to overcome sampling issues.
  • Particularly during eSurveys, user testing to establish whether the responses reflect an understanding of the questions as intended is critical.
  • As with all pilot projects, issues remain concerning the technology. UMCP protocols are unsupported in certain locations and are unfamiliar to many users. Further exploration of user behaviour is essential.