Gender and Network Formation in Rural Nicaragua: A Village case study

Abstract

This contribution examines the relation between gender and network formation in rural Nicaragua in 2007 and studies differences in the structure and contents of men's and women's networks. Such differences are relevant, as network theory suggests that structural characteristics – as well as the contents of networks – strongly influence the type and amount of benefits generated. Through the application of dyadic regression techniques, this study examines the determinants of the size and socioeconomic heterogeneity of individual networks. Research findings suggest gender segregation of networks and considerable differences in the structure and content of men's and women's networks. These differences relate to the gendered division of labor and to women's time poverty in particular. Our results are relevant in a context where policy makers increasingly consider social networks an important policy tool. We caution against a gender-blind alignment on existing social networks and argue for detailed mapping and unpacking of social networks through a gender lens.

Keywords:

• Social network analysis
• dyadic regression
• gender sorting
• gendered labor division
• rural Nicaragua

Acknowledgments

We wish to thank the Nicaraguan research and development institute Nitlapan-UCA for its cooperation and support during the fieldwork. We are also indebted to Miguel Alemán, Guy Delmelle, Selmira Flores, Ligia Gómez, Francisco Pérez, Arno Riedl, and Alfredo Ruíz for interesting methodological discussions; and Tania Paz Mena, Leonardo Matute, Francisco Paiz Salgado, Edna García Flores, Fátima Guevara, Silvia Martinez Arróliga, and Will Tellez for research support in the field. We are also grateful to Francisco Varela, a local community leader, who supported our fieldwork. Lastly, we wish to gratefully acknowledge the financial support of the Flemish Institutional University Cooperation (VLIR-UOS) and the Institute of Development Policy and Management at the University of Antwerp.

All personal information that would allow the identification of any person or person(s) described in the article has been removed.

Notes

¹ Within the area of demographic research, for instance, there is a strand of literature that studies the importance of differences in structure and content of (conversational) networks for contraceptive use in developing countries (see, for example, Hans-Peter Kohler, Jere R. Behrman, and Susan C. Watkins 2001).

² The SIGI is a new composite measure based on variables of the OECD Gender, Institutions, and Development database (Boris Branisa, Stephan Klasen, and Maria Ziegler 2009). It combines five dimensions of social institutions related to gender inequality (Family Code, Civil Liberties, Physical Integrity, Son Preference, and Ownership Rights).

³ Other useful sources include Popielarz (1999); McPherson, Smith-Lovin, and Cook (2001); Liebler and Sandefur (2002).

⁴ See also Rankin (2002); Cagla Okten and Una Okonkwo Osili (2004); Bandana Purkayastha and Mangala Subramaniam (2004); Arganosa-Matienzo (2005); and Cleaver (2005).

⁵ See Mandel (2006) for a discussion of the constraints on women's mobility in the Global South and, more generally, an analysis of the centrality of spatial processes to women's economic and social empowerment.

⁶ This term does not make any judgment about who takes the lead in household decision making. None of the households consists of same-sex or nonmarital relationships.

⁷ Because some households or household nodes were not available at the time of the interview, we were unable to include all sixty-six households or interview all 123 household nodes.

⁸ This difference in perception may have several reasons. It may, for instance, be the result of people putting a different emphasis on the different “contents” of relations (such as friendship relations, neighbor relations, or economic relations).

⁹ This refers to the inclusive OR condition used to define whether a tie exists. The alternative would have been to use so-called AND-networks where ties are taken to be valid only if A “and” B mention the relation. There are several arguments, however, in favor of the use of OR-networks, which are related to the risk of excluding existing ties when using AND-networks. First, those who have many ties are more likely to forget to mention a tie than those who only have a few ties. Second, people place a different emphasis on relations of different contents, which may lead to people excluding relations of certain contents. Third, there may be asymmetry in reporting the direction of benefits provided (for example between receiving support and giving support).

¹⁰ Total possible undirected ties among women: 1,891 = (62 × 61)/2; possible undirected ties among men: 1,830 = (61 × 60)/2; possible undirected mixed ties: 3,724. The last result is derived from subtracting 1,891, 1,830, and 58 (the number of possible undirected intrahousehold ties) from 7,503 = (123 × 122)/2 (that is, the total number of undirected ties between all nodes).

¹¹ As we have sixty-one men and sixty-two women, the maximum number of ties for a male node with other men is sixty, for a female node with other women sixty-one, for a male node with women sixty-two, and for a female node with men sixty-one.

¹² One could argue that endogeneity may lead to biased estimates of the coefficients of the sum and difference on contact urban center. First, endogeneity may be the result of omitted variable bias: people who are more likely to go to urban centers could have some unobserved characteristics that are correlated with the probability of network formation. In rural Nicaragua, however, most people only visit urban centers if necessary for their economic activities and if they have the economic means to do so. In other words, the strongest determinants of contact urban center are economic activities and wealth, which we control for by including land and cattle variables. Second, endogeneity may also be the result of inverse causality. For example, people who have a social tie may adapt their daily activities to each other, which may reduce differences on time spent in the urban center. However, here again we argue that individual decisions to visit the urban center (and, consequently, the sum and differences on it) are strongly dependent on economic activities and wealth, and thus the influence of social networks is limited. Moreover, the possibility of reverse causality is limited by the fact that we define the variable contact urban center as the number of visits to the urban center in the previous month, whereas the social network data refers to the current period (meaning the moment of interview).

¹³ As the functioning of village committees, cooperatives, and local representations of political parties requires interaction with higher administrative levels, we expect education to be important for those men who lead these social public activities (on this topic see, for example, Ben D'Exelle 2009) but not necessarily for all men who participate in these activities. This situation explains why differences in years of education exert a statistically significant effect while the sum of the years of education does not.

¹⁴ This finding does not imply that women's mobility is not an issue. To control for this, we should have included a variable that measures the geographical distance to each public spot, but this is information that, unfortunately, we did not capture.