The Mumbai-based Bharatiya Muslim Mahila Andolan, an advocacy group for Muslim women, published a survey in 2015 which among other things found that 92.1 per cent of Muslim women wanted a ban on triple talaq, the oral unilateral divorce that Muslim personal law currently allows men to use to divorce their wives. By any standard of credible data though, neither that number nor the rest of the survey are of much use in the national debate.
The Mumbai-based Bharatiya Muslim Mahila Andolan published a survey in 2015 which among other things found that 92.1 per cent of Muslim women wanted a ban on triple talaq.
The first and most important feature of a survey is whether its sample is representative of the universe it is drawn from or seeks to speak of, in this case all Indian Muslim women. In every possible way, the BMMA survey does not fit the bill.
This should have been clear by looking at the sample, which the BMMA describes in its report. The BMMA interviewed 4710 women in ten states. More than 35 per cent of respondents were from Ahmedabad (Gujarat), Dindigul (Tamil Nadu) and Cuttack (Odisha). The actual share of these three cities to the country's female Muslim population is 0.65 per cent put together. What pollsters would do to smooth such issues out would be to weight the responses proportionately (although this would be too big an issue to smooth); the BMMA did not.
Secondly, 11.1 per cent of the adult women they surveyed were divorced, while Census 2011 data shows that even while the rates of divorce among Muslim women are higher than those for women from other religious communities, just 0.4 per cent of adult Muslim women are divorced.
The sample of women the BMMA surveyed was atypical, and not representative of Indian Muslim women. And there's a very good reason why this happened.
All of which goes to say that the sample of women the BMMA surveyed was atypical, and not representative of Indian Muslim women. And there's a very good reason why this happened.
When research organisations like the National Council for Applied Economic Research or the Centre for the Study of Developing Societies begin a study, they create a 'sampling frame', usually based on the Census and the National Sample Survey, and draw up a sample that is as geographically and demographically diverse as the country is. Any remaining blips are fixed through weights.
I asked Noorjehan Safia Niaz, the report's lead author and a respected voice for Muslim women, how the sample was drawn. The BMMA contacted women they knew or women from communities in which they already worked. They surveyed these women and then honestly described their sample.
The BMMA contacted women they knew or women from communities in which they already worked.
This is a frequent problem with surveys conducted by advocacy groups - polling the atypical communities they work with instead of representative samples. Some older, better quality polling of Muslim women's attitudes to marriage and divorce do exist. Sixty per cent Muslim women told CSDS in its 2006 State of the Nation survey (Muslim women all-India sample size: 540) that they fully agreed with the statement that triple talaq at one go was wrong (quite different from wanting a ban). An additional 13.7 per cent said that they "somewhat agreed" that triple talaq was wrong. An even larger proportion was opposed to polygamy. But over half (52 per cent) of Muslim women respondents in their 2008 State of the Nation study said that even if a husband and wife were incompatible, they must stay together.
The CSDS numbers too point to Muslim women's opposition to triple talaq, though not at the scale currently being reported. This doesn't necessarily change the contours of the debate; but it should at the very least change the way survey data is reported.