In this dissertation, three primary concerns about the limits of poverty and generosity are addressed. First, from a methodological point of analysis, we examine the variations in the application of the Factorial Survey method found in these three studies. Second, a more in depth examination of attitudes about, and conceptions of, the poor and the lower limits of poverty is conducted using one of these studies, the 1986 General Social Survey. Finally, a comparison of public attitudes and prevailing public policy is used as a backdrop for the discussion of the social policy implications the findings from these studies offer. The most prominent finding presented in this dissertation is that there appears to be a monetary barrier, or "floor", below which a vast majority of the American public believes no one should be allowed to fall. Over 90 percent of the respondents averaged awarding $200 or more per week to the vignette families, regardless of the circumstances presented in the scenario. Less than 2 percent of the respondents averaged less than \$150 per week in award. Second, this floor represents over TWICE the average payments that are actually provided by the Government to those persons who are deemed eligible for assistance. Indeed, for a majority of the regions in this country, the average income awarded to the vignette families was over 3 times the AFDC awards given to recipients in the lowest supported state for that region. Overall, in the analysis of the Vignette Supplement to the 1986 General Social Survey, the American public perceives that poor families with children, in almost all circumstances, are deserving of support, and that the level of support needed, and approved, is twice the amount currently provided by government programs.

This dissertation examines the gender differences among the homeless. Although women with dependent children comprise the fastest growing subgroup among the homeless population, most of the homeless are still single adults. The vast majority of these single adults are men, though the number of single women has increased notably in recent years. The second purpose of the dissertation is to examine who is most vulnerable to homelessness. The analysis strategy was to examine the literal homeless comparing three subgroups, unattached men, unattached women, and women with dependent children, and then to examine the very poor who are housed comparing the corresponding subgroups. Comparisons were made of the characteristics of and the financial and social resources available to the literal homeless and the domiciled poor. The descriptive analysis of the gender differences among the homeless are based on empirical data collected from the Chicago Homeless Study in 1985. Data on unattached men and women who are housed come from the General Assistance Study conducted in Chicago in 1984 and data on women with dependent children come from a 1985 study of welfare families in Chicago. The findings indicate that homeless men are poorer, more disabled, and more socially isolated than homeless women. In particular, their levels of poverty, disability, and social isolation are substantially higher than that of homeless women with dependent children. Their high levels of disabilities make it extremely difficult for the homeless men to find employment or to reciprocate in social relations. Most of them have had their social networks destroyed or exhausted and few receive needed social services. The absence of social and public resources means that homeless men have difficulty extricating themselves from homelessness. For these reasons, the men are more vulnerable to chronic homelessness. There is a subset of homeless single women whose conditions are not very different from homeless men. Women with dependent children have the best chances of extricating themselves from homelessness. Policy implications are discussed.

This dissertation analyses the survival strategies that wealthy people in Mexico City have designed and implemented to protect themselves from kidnapping with special focus on household employment relationships. This particular crime has demonstrated a particular evolution in the last 20 years that deserves analysis. Once a political crime, it became an economic crime that at first only targeted wealthy individuals and then over time began targeting working class victims. Based on extensive qualitative fieldwork in Mexico City which included a year in the field, 78 interviews with employers, employees, kidnapping victims and members of the police forces and justice system and the creation of a news reports database this thesis presents a detailed history of the evolution of kidnapping in the period 1968-2009. This is followed by an in depth analysis of the strategies elites use to protect themselves from this crime. Special attention is focused on the hiring process of household employees, namely drivers, as evidence suggests that most kidnappings are organized or facilitated in some way by a close collaborator of the victim. The hiring process is approached as a problem of trust. Signaling theory is the main framework used for the solving of this problem, as well as some ideas found in transaction cost economics, namely vertical integration. The results point towards strategic behavior from the actors involved that seeks to minimize the risk of being kidnapped for the employer. Signaling helps us uncover the specific mechanisms by which employer establish their prospective employees’ trustworthiness. The use of informal social networks made up of strong ties is one of the most salient mechanisms used to guarantee honest employees and this, together with a composite set of properties is signaled throughout. This thesis contributes to the literature on crime in Latin America as well as to the sociological literature on signaling, a branch of analytical sociology.