173 - Building a "Voicebook" of the Voiceless for Large-Scale Real-Time Program Evaluation

Abstract Information: Interactive voice response (IVR) is an automated phone system that greets callers and allows them to access pre-recorded audio information and acts on their responses. Most people in developing countries such as the United States associate the IVR system with telemarketing and customer services. However, it has been adopted in many developing countries to reach underserved population groups such as rural villagers and migrant workers, providing assistance for them to access social services and healthcare. In this presentation, I will illustrate how the IVR system can be purposefully incorporated into the program design, implementation, and evaluation by building an application akin to a “voicebook” to not only include but also amplify the voice of the voiceless for health promotion. Main Kuch Bhi Kar Sakti Hoon (“I, A Woman, Can Achieve Anything”) is a multimedia initiative led by the Population Foundation of India along with NGOs and community partners to challenge regressive gender norms and advocate for the health and empowerment of women and girls in rural India. This was accomplished through the story of its protagonist, Dr. Sneha Mathur, and her trials and tribulations through three seasons of serial drama broadcast on the Indian national network Doordarshan and 216 All India Radio stations. A community-based participatory technology company Gram Vaani (“Voice of the Village”) pioneered the IVR system model with the Population Foundation of India to leverage India’s vast mobile telephony infrastructure to reach rural residents, empower local communities, and enable prosocial actions. Across three seasons, a toll-free number was advertised during the broadcast that audience could give a “missed” call to receive a free callback and participate in different activities by pressing numeric keys and recording voice messages. This helped us collect call log data and information based on caller participation such as surveys, quizzes, and discussions. A total of 2,137,049 calls came from 430,071 unique callers. The survey data suggested that participants came from 36 states, especially the northern states as the project priority locations. Those who answered weekly quizzes for a prize showed an average of 82.4% correct rate (ranging from 50.6% to 97.6%). Audience voice messages were routinely monitored and tagged with keywords relevant to the project’s health themes and prosocial messaging. In their own words, callers discussed what they thought about the program in general, the storylines and main characters, embedded educational issues, and their personal opinions and stories of transformation. On average, these messages were about 30 seconds, with some as long as 2 minutes. After screening, most of them were “published” on the Gram Vaani mobile network and made available for other callers to listen in, which in turn encouraged and inspired more calls and voice messages. In fact, this large-scale real-time IVR system was so effective that the program was able to adjust broadcasting schedule and tailor their content in subsequent seasons to feature youth and address the role of men in women’s health and empowerment based on audience suggestions.