The Quarterly Journal of Economics, 2009, vol.124, iss. 3, pp. 1057-1094. Available From:Link to Source
This study investigates the effects of cable television access on women’s status in India, using individual-level panel data from five states (Bihar, Delhi, Goa, Haryana and Tamil Nadu) from 2001 to 2003. A great deal of popular cable programming departs from traditional India in terms of lifestyle, especially for women. For instance, female characters typically have fewer children, receive more education, are more independent and often work outside of the home. Whether being exposed to such information is enough to change attitudes and behaviours is the empirical question the authors attempt to answer here.
Because the assignment of cable television is clearly non-random, with remote villages less likely to receive coverage, the authors adopt a difference-in-differences identification strategy. This allows for ‘treatment’ and ‘control’ villages to be different along certain dimensions, as long as their outcomes follow parallel trends.
The study period witnessed a rapid expansion of cable television in India. Of the 180 sample villages, 21 received coverage for the first time between 2001 and 2003. Key outcomes being studied are attitudes towards domestic violence and son preference, measures of female autonomy and pregnancy as well as school enrolment for girls. For each outcome, the authors plot the values in 2001, 2002 and 2003 for four separate categories of villages (those with cable in 2001, those receiving cable in 2002, those receiving cable in 2003 and those still without cable in 2003). Doing so allows readers to visualise the magnitude and timing of effects. This is complemented by results from regressions, in which an outcome (e.g., son preference) is regressed on an indicator of cable coverage, individual-level and village-level controls.
The authors find substantial effects of cable television access on numerous survey measures of women’s status in India. One such measure is the number of situations in which physical beating is reported to be acceptable by women. Amongst six hypothetical situations (e.g., wife being unfaithful, wife’s family doesn’t give money, wife doesn’t take care of children), those receiving cable for the first time in 2002 show a drop from an average of 2.2 acceptable situations in 2001 to 1.9 in 2002 and to 1.75 in 2003. By contrast, those having cable before 2001 and those without cable by 2003 (the control groups) show no difference in this outcome throughout 2001 to 2003.
A very similar pattern is observed for son preference. For the group that received cable for the first time in 2002, 55 per cent reported wanting their next child to be a boy in 2001, compared with only 45 per cent in 2002. The group that received cable for the first time in 2003 shows an even larger drop from 59 per cent to 45 per cent from 2002 to 2003. Similar effects are also found for measures of female autonomy. Furthermore, cable television access appears to be correlated with a decrease of 4 percentage points in the likelihood of pregnancy. In terms of school enrolment, there is evidence of a positive effect (roughly 5 percentage points, from a baseline of 65 per cent) on girls aged 6 to 10 years, whilst boys and girls aged 11 to 14 years do not seem to benefit. Regression estimates show that these effects are significant at the 5 per cent level. To sum up, cable television appears to be very effective at changing traditional attitudes and behaviours in rural India, suggesting a potentially novel intervention for improving women’s status.