Srinivasan Vasudevan

Can a short program create friendships among isolated women? We randomly invited 84 of 166 young married women in peri-urban Gujarat, India, to six facilitated sessions over 12 weeks. At program completion, invited women reported 0.8 more friends. Eighteen months later, invited contact had become chosen contact: invited women met more invited and fewer control women, and shared invitees made control women likelier to befriend each other. About 22 months later, their choices in a coordination task matched those of 1.4 more women; information collection did not detectably change. It cost under USD 55 per attending woman and USD 90 per new friendship.

Do India’s elite engineering institutes raise the innovative output of their graduates? I link the records of all candidates in the 2009 Joint Entrance Examination, including their scores and school-leaving marks, to institute rosters and LinkedIn profiles for enrollment, and to patent filings and OpenAlex publication records for output over the following fifteen years. I estimate effects using regression discontinuity designs at three admission cutoffs: eligibility for an Indian Institute of Technology, admission to an established rather than a new one, and admission to a newer rather than an older field of study.

How did the arrival of cable television change rural Indian households? The data come from the 1993–94 Human Development Profile of India and the 2004–05 and 2011–12 rounds of the India Human Development Survey, which record when cable reached each village along with media consumption, social participation, and norms and beliefs. I estimate staggered difference-in-differences comparing villages that received cable with villages that had not yet adopted it or never did.

Does supporting grassroots journalists improve local reporting? Local news in India is dominated by sensational stories, while environmental problems that hit rural communities hardest, such as drought, flooding, and heat stress, go largely unreported. Newspapers rely on stringers, part-time local correspondents paid little per story. We ask whether training and supporting stringers increases the quantity and quality of local reporting on climate and environmental problems, and whether better local coverage in turn changes political, policy, and socioeconomic outcomes in the communities they cover.

Can state-run mass media build a nation out of a linguistically and ethnically diverse population? After independence, India expanded All India Radio (AIR), a government monopoly that carried news and unity-promoting programming, in a staggered rollout under the five-year plans. We assemble the stations’ primary service areas from archival broadcasting records and link them to districts. Outcomes come from the linguistic tables of the 1931 and 1961 censuses and from vote shares of national and regional parties in constituency results of the 1951–52 and 1957 general elections. We estimate the impact of radio on national cohesion using staggered difference-in-differences comparing districts that gained radio coverage with those that gained it later or never gained it.

During the 2014 Indian general elections, we carried out a large-scale experiment randomizing a radio campaign highlighting the disadvantages of voting for corrupt vote-buying candidates. Official electoral data show that the radio campaign significantly decreased the vote share of parties that engaged in the most vote buying (as reported by journalists). Voter survey data show that the campaign increased the salience of corruption as an election issue and decreased voting for parties that offered gifts. From a policy perspective, we show that radio campaigns are a cost-effective method to influence voter behavior. From a measurement perspective, we show that journalist interviews can impartially identify vote-buying parties.

Can mass communication play a role in spurring technology adoption and productivity growth in agriculture, particularly when there are significant barriers to social learning? To answer this question, I estimate the impact of regional radio broadcasts that included significant farm programming, on the adoption of high-yielding varieties (HYVs) of rice and wheat during the Green Revolution in India. I exploit the staggered expansion of the regional radio network to identify the impact. Using an annual district-level administrative dataset from 1966 to 1978, I find that introduction of regional farm radio broadcasts led to a large persistent increase in the adoption of HYVs of rice, a spatially heterogeneous technology less amenable to social learning. By contrast, there was at most a transient impact on the adoption of HYVs of wheat, a spatially homogeneous technology with a lot of scope for social learning. Importantly, I find that the regional farm radio broadcasts increased rice yields by 15%, a magnitude potentially large enough to justify the investments in the expansion of the regional radio network. Overall, this paper demonstrates that mass communication can indeed play a key role in diffusing profitable agricultural innovations when there are significant barriers to social learning.