Congratulations to the 2008-2009 Seed Grant Award Recipients! Scroll down to see all four project abstracts.
Improving Recall of Survival and Mortality Events during Retrospective Interviews in Developing Countries: A Randomized Controlled Trial of Supplementary Interviewing Techniques
Stéphane Helleringer (Department of Population and Family Health)
James Phillips (Department of Population and Family Health)
Reliable estimates of mortality are essential for the planning and evaluation of health interventions in resource-poor settings, yet very few developing countries have complete vital registration systems. Instead, mortality rates are estimated from retrospective data collected during household-based surveys, by asking respondents about the survival of close relatives (e.g., children, siblings) (e.g., Timaeus and Jasseh, 2004). While such indirect estimates are inexpensive and have produced valuable information on child survival and adult/maternal mortality, the estimates produced often underestimate mortality (Gakidou et al., 2004; Smith et al., 2001). This is the case because 1) indirect estimation methods rely on strong assumptions that are often violated in practice, and 2) the number and timing of deaths are frequently underreported during retrospective surveys. The main aim of this project is thus to develop and test new supplementary interviewing techniques designed to improve recall of vital events during survey interviews.
Our basic strategy towards this goal consists of identifying the individual-level characteristics of both respondents and their relatives that predict omission/misreporting of deaths during retrospective surveys. In this context, the specific aim of this CPRC seed grant application is to explore the feasibility of linking individual records of deaths collected prospectively to retrospective reports of siblings’ and children’s mortality collected during a household survey. We propose to conduct 200 retrospective interviews in the Kassena-Nankana district of Northern Ghana (Navrongo DSS) to determine what proportion of retrospective reports are successfully linked to prospective reports of deaths. If successful, the data collected during this seed grant will provide the preliminary data necessary for the development of an ambitious R01 application to the NIH. The proposed project represents a critical building block in the development of a formal collaboration between centers and departments involved in demographic research at Columbia University and universities/research sites in Ghana, Tanzania and Senegal.
We Can Be Good Parents: Understanding and Mitigating Reproductive Stigma Against HIV-positive Men and Women
Jennifer Hirsch (Department of Sociomedical Sciences)
Wafaa El-Sadr (Clinical Medicine and Epidemiology)
Jessica Justman (Clinical Medicine and Clinical Epidemiology)
Nora J. Kenworthy (Department of Sociomedical Sciences)
We propose preliminary community- and clinic-based research on reproductive stigma among HIV-positive individuals and couples living in two HIV-endemic countries, Kenya and Lesotho. This pilot project is the initial step towards a larger, multi-site program of intervention research to understand and mitigate the effects of reproductive stigma. This program of intervention research will seek to alter community and provider norms about the social value of reproduction for HIV positive individuals. The intervention would have a clinical health objective—to improve health outcomes for HIV-positive individuals and their partners—as well as a population-level health objective: to reduce stigma in order to enhance uptake of HIV testing and facilitate earlier entrance into care. The proposed pilot research will be essential to crafting a successful NIH grant proposal for a comprehensive program of intervention and evaluation in Lesotho and Kenya. Drawing on the theoretical concepts of stratified reproduction and reproductive justice, the pilot project aims to: 1) conduct preliminary research (key informant interviews, focus groups, semi-structured interviews, and discourse analysis) on stigma and the social value of reproduction with key stakeholders in two communities in which existing ICAPsupported treatment and care programs provide considerable scientific and logistical infrastructure for the development of future interventions; 2) assess the appropriateness of each site for the subsequent development of an intervention to mitigate stigma through affirmatively valuing and enabling the reproduction of HIV positive individuals; 3) analyze the qualitative data collected in each site, producing a manuscript for publication and an R01 for submission to NIH.
The proposed pilot research meets the aims of the CPRC seed grant program: it would develop cross-departmental partnerships between SMS and ICAP; it would develop mentorship and collaboration among investigators at different career stages (ICAP directors, Associate Professor, doctoral student / CPRC fellow) as well as with colleagues Lesotho and Kenya; it would bring social science theory to bear on the development of a population-level intervention for a critical health issue; it would build on existing research infrastructures and capacities in two exceptional ICAP country sites; it would bring interdisciplinary social science research initiatives to already existing medical and health systems interventions at those sites; and it would strongly enhance the development and fundability of a future NIH grant to support implementation and evaluation of a broader research project.
Monitoring Public Opinion of Immigrants in the U.S.
Neeraj Kaushal (School of Social Work)
Francisco L. Rivera-Batiz (Teachers College)
The objective of this proposal is to develop research to study the correlates of public attitudes towards immigrants. The proposed seed grant will be used to develop a research proposal that will be submitted to the National Science Foundation to study: (i) changes in public attitudes towards immigrants and immigration policies since the mid-1990s, (ii) the factors that determine these attitudes, (iii) the likely impact that changing public perceptions may have on existing policies relating to immigration and immigrant integration, and (iv) shifts in attitudes towards undocumented immigrants and differences in public opinion towards legal and illegal immigration. Preliminary research will be conducted using the 1994 and 2004 General Social Survey of the NSF on public attitudes towards immigrants. The findings from this research will be included in the grant application to the NSF to elaborate its potential contributions and merits.
Latent Structure Models for Social Networks Using Aggregated Relational Data
Tian Zheng (Statistics)
Tyler McCormick (Statistics)
Questions of the form "How many X's do you know?" collect aggregated relational data from one's social network and are a common means of learning about populations that are hard to reach or survey directly. McCarty et al. (2001), for example, take X to be individuals who are HIV positive, are homeless, or were recently raped to estimate the size of these traditionally hard-to-count populations. The United Nations Development Programme (UNDP) also currently funds multiple projects using aggregated relational data to estimate the size of key populations at risk of developing HIV/AIDS with the hopes of using these estimates to more effciently allocate HIV/AIDS prevention resources.
We propose a method to estimate more complicated structural properties, such as clustering, in networks using aggregated relational data. Our method is similar in spirit to the latent space models proposed by Hoff and Handcock (2002) when the entire network is observable and builds on recent methods to learn about more basic features of social structure from aggregated relational data by Zheng et al. (2006) and McCormick et al. (2008).
Our method makes information about more complicated network structure available to the multitude of researchers who cannot practically or financially collect data from the entire network, such as researchers and policymakers interested in estimating the size of hard-to-count populations or in learning about how these populations interact with other groups. Our models can also be used to generate null distributions for empirically quantifying uncertainty about estimates of particular network features.