This is a joint post with Frances Simpson Allen, of the UN Millennium Campaign
Women doing the MY World survey in Bangladesh
The million votes – and counting – in the MY World survey will keep data geeks (like us) happy for years. There’s lots of stories in there, but there’s one in particular which has us scratching our heads and we’d like to know what others think. It’s this.
We have been experimenting with different ways of collecting data via mobile phones. About a fifth of the votes come in via phone, mainly text messages or people calling a number and recording their answers. And while overall the votes are pretty evenly balanced between men and women, the phone votes are strikingly male (all the data can be found here).
Globally, two men answer the survey by phone for every one woman. Of the fifteen countries where there are more than 2000 mobile phone votes, there were only two, Nicaragua and Kyrgyzstan, where women’s votes outnumbered men’s. In India, Bangladesh and Ethiopia more than 90 per cent of phone votes come from men. In another five countries: Nigeria, Zimbabwe, Kenya, Pakistan and Nepal, more than three quarters of phone votes come from men, while in Ghana, the DRC and South Africa, between 65 and 75 per cent of mobile votes came from men. India and Nigeria also have large numbers of online votes, and there men also outnumber women though by a much smaller margin (around 60:40).
This imbalance is much higher than rates of phone ownership would suggest. According to industry research, across Africa, 56 per cent of mobile phones are owned by men, and 44 per cent by women, while in South Asia the figures are 63 per cent men and 37 per cent women. The Kenya figure is also particularly striking, given that a recent survey suggests there is very little difference in mobile phone usage rates between men and women.
The SMS MY World survey so far is not sampled (although we are doing sampled surveys through other means in a growing number of countries) and relies on people choosing to take part. It’s clear that, under these conditions, women are less likely to offer information via mobile. With one of our mobile partners we are testing female-specific messaging to see if more women opt-in when the question is phrased differently. We don’t have results from this pilot yet and we are keen to find out from others involved in similar exercises if the gender imbalance is a general finding or something specific to MY World.
There is huge optimism about the possibilities of using current technology to improve data collection, and through that to improved transparency and accountability. Much of that is justified. But the old problems – making sure that data is representative and doesn’t reflect existing biases and inequalities – apply to new technologies as much as to more traditional ways of collecting data. The experience with MY World suggests that this may be a bit harder than we’d like to think.