Linear Models Methods for Some Rank Function Analyses of Ordinal Categorical Data

This paper is concerned with the analysis of ordinal data through linear models for rank function measures.Primary attention is directed at pairwise Mann-Whitney statistics for which dimension reduction is managed by use of a Bradley-Terry log-linear structure.The nature of linear models for such quantities is contrasted with that for mean ranks (or ridits).Aspects of application are illustrated with an example for which results of other methods are also given.     

Exploring Moderating Effects of Violent Conflict on Household Welfare in a Resource-Constrained Environment: The Case of Sub-Saharan Africa

ABSTRACT

This study examines correlates of household welfare in three countries of sub-Saharan Africa, Tanzania, Zambia, and Zimbabwe, using data from the 2004 wave of the Afrobarometer survey. We also assess the role violent conflict might play in influencing this relationship. This subject has become especially relevant today in sub-Saharan Africa, given the growing disfranchisement of vulnerable individuals and households and increased incidents of violent conflict. Insight into the relationship between violent conflict and household welfare may be an important step in understanding why many sub-Saharan African countries have difficulties in stimulating economic growth and welfare. The study sample includes 3,525 respondents. Study findings provide partial support for the hypothesized relationship. Specifically, poverty reduction initiatives and informal assistance are associated with reduced hardship, while violent conflict is related to an increase in hardship. We also note that certain individual and household characteristics are linked to hardship. Practice and policy implications are suggested.     

Social Class Differentials During Fertility Decline: England and Wales Revisited

This paper reviews the relationship between landholdings and fertility. Two dimensions of land are identified as salient for fertility behaviour: size of operational holdings and land ownership. It is suggested that these two dimensions and the resulting income streams have disparate effects on fertility. Size of holdings is assumed to have a positive influence on fertility due to the greater labour demands of larger holdings, while land ownership is posited to exert a negative long-term effect because of the increase in old-age security associated with the income returns to equity. In addition to these effects on the demand for children, landholding is also thought to influence the supply of children. A systematic review of the literature finds support for the impact of both dimensions of landholding on fertility preferences, contraceptive behaviour, the proximate determinants and fertility. Both the demand and supply of children appear to be influenced by landholdings. The observed regularities suggest the need for further research on this connection, not the abandonment of this line of inquiry.     

Assessing Attitudes Toward Lesbians and Gay Men: The Modern Homophobia Scale

The authors developed the Modern Homophobia Scale (MHS), a scale that includes measures of attitudes toward lesbians (MHS-L) and attitudes toward gay men (MHS-G). An initial pool of 97 items was generated by the authors based on a review of the literature on homophobia and feedback from several focus groups. Responses to the initial pool of items were factor analyzed. The final versions of the MHS-L and the MHS-G included items reflecting the following factors: Personal Discomfort with lesbians/gay men, Institutional Homophobia toward lesbians/gay men, and the Deviance/Changeability of female/male homosexuality. Although the Personal Discomfort, Institutional, and Deviance/Changeability factors were separable, they were highly correlated and showed little evidence of divergent validity. Reactions to gay, lesbian, and heterosexual characters in written vignettes provided initial support for the validity of the MHS. Both the MHS-L and MHS-G were significantly correlated with the Index of Homophobia and the Attitudes Toward Women Scale. A greater degree of homophobia on the MHS-L and MHS-G was associated with a greater degree of socially desirable responding as measured by the Balanced Inventory of Desirable Responding. Lower levels of homophobia on the MHS-L and MHS-G were associated with having lesbian and gay acquaintances and friends. On the MHS-L, men and women did not differ in their overall level of homophobia toward lesbians, but significant gender differences on MHS-L factor scores were present. Compared to men, women were more supportive of gay men on all three factors of the MHS-G.     

The effect of female migration on infant and child survival in Uganda

This article uses data from the 1996 Uganda Demographic and Health Survey to examine whether migration of women improves the survival chances of their children to age five. We expand on prior research by testing not only the hypothesized positive effect of rural-urban migration, but also the effects of other migration stream behaviours on child survival. Results show that up to 10% of children die before age five and within-group differences in mortality exist among urban and rural children depending on their mother's migration status. Only urban-urban migration was significantly related to child survival, compared to rural non-migrants, after controlling for other factors, although other streams of migration (rural-urban, urban-rural, rural-rural) were positively related to child survival. Generally, migration explains a small component of the variance in child survival. Several other factors, including parents' education, household size, household headship, mother's age at birth, duration of breastfeeding, and place of delivery have a significant predictive power on child survival.     

A nonprofit manager's guide to online volunteering

As access to the Internet expands, volunteer managers have the option of using this new technology to manage pools of volunteers online. Online volunteering can appear deceptively simple: post an opening, select a volunteer, and then manage them by e‐mail. However, our evaluation of an active online volunteering service used by more than three hundred organizations and ten thousand potential volunteers indicates that additional planning and management are frequently needed. Through an analysis of two years of quantitative data and interviews with more than forty volunteers and managers, we have extracted suggested practices and programmatic implications for managers in the areas of program design, volunteer selection, training, ongoing operations, and evaluation.