The majority of sociological research on job search strategies has focused primarily on social networks with little attention given to all other forms of job search. Also, much of the existing literature focuses on outcomes of a job search as opposed to the search itself. This paper seeks to expand the job search literature by focusing on the characteristics that determine which job search method an individual will use and then predicting job retention as a result of the job search method. Using data from the Louisiana Job Search Survey (2002) I find that network structure has an effect on choosing personal contacts as a job search method. Particularly, having a higher proportion of weak ties in the network leads to higher probabilities of using personal contacts. I also find that job search methods vary by metropolitan and nonmetropolitan areas. I do not find that the job search method a person uses has an effect on job retention. The overall findings suggest that job search is not determined by personal characteristics, but instead seems to be related to the situations that surround job searches.

Evidence suggests that social support can mitigate some of the harmful effects of stress on health. Social support theorists argue that certain social groups have differential access to social support; therefore, certain social groups are at a higher risk of experiencing psychiatric symptoms. Although social networks are beyond the scope of these analyses, it is an important component to consider when examining the uneven distributions of social support between social groups. If racial differences exist in the networks in which individuals are embedded, then part of the differential access to social support could be explained by examining the various compositions of networks. This thesis examines racial differences in the social support process. Using data collected on recovery support during Hurricane Georges, I examine: (1) whether and how social support systems for blacks and whites differ in nonroutine situations and (2) whether the relationship among stress, support, and depression differ between blacks and whites. My results indicate that blacks are less likely than whites to receive instrumental support in the preparation phase of the storm and blacks who receive less instrumental support after the storm are more likely to suffer from depression. This suggests that future research should explore systematic differences in blacks and whites network characteristics, including the types of resources, and the variations in the network structure.

This study investigates support for the death penalty and federal crime spending among white Americans. Data from the American National Election Studies (ANES) series (1992-2000) are matched with census tract level indicators of demographic and community characteristics from the 1990 and 2000 Census Bureau Summary File Tape 3A and county level crime data supplied by the Federal Bureau of Investigation�s Uniform Crime Reports (UCR). Ordered logistic regression is used to investigate five general research questions: (1) Are racial attitudes the most salient individual level predictors of support for crime policy among whites as suggested by prior research? (2) Are whites� crime policy preferences influenced by the structural environment? (3) Is the relationship between key individual level variables (i.e. racial attitudes) and support for crime policy moderated by features of the community? (4) Does fear of crime influence white support for crime policy? (5) Are there different explanatory models of white support for the death penalty and federal crime spending? The results of the study show whites� racial attitudes are strong predictors of white support for the death penalty and federal crime spending. Overall, structural conditions do not shape white support for the death penalty. On the other hand, whites� crime spending preferences are influenced by the racial context of the community. Whites living in more racially segregated communities are less supportive of increased crime spending. Turning to the interaction results, the relationship between whites� racial attitudes and support for the death penalty is mediated by features of the community (i.e. racial composition, segregation, deprivation). In general, interaction effects are not present in the crime spending models. The fear of crime models show fear is related to white support of crime spending but is not associated with support for the death penalty. In sum, these findings suggest further investigation of white support for crime policy is warranted.

Most large-scale studies on voluntary childlessness since the waning of the baby boom provide cross-sectional estimates for a single time period. They cannot be synthesized to estimate change because of the varied definitions used to operationalize voluntary childlessness. In this study, I use data from the 1973-2002 cycles of the National Survey of Family Growth (NSFG) to estimate change in voluntary childlessness using a consistent definition by period and birth cohort. I find that voluntary childlessness stayed relatively constant through the seventies and eighties, but showed a large increase from the mid-nineties to 2002. I show that voluntary childlessness increased in recent years because baby-boomers postponed childbearing until they no longer desired it, and younger women born in the seventies are now deciding to remain childless earlier. I discuss the role of these younger women in establishing a ceiling for voluntary childlessness. I also provide initial results supporting the theory that voluntary childlessness is diffusing among women of lower education and higher religiosity.

In the past, much of the literature on marital quality and stability has suggested that effective communication, positive affect, and spouses emotional support are important to marital quality (Gottman and Levenson, 2000; Gottman and Notarius, 2000 Popenoe, 2004; Sanchez and Ganger, 2000; Whitehead and Popenoe, 2001). Other works have also suggested that commitment, that is feelings regarding the permanency of marriage, encourages marital quality. However, to date, few studies have empirically investigated the relationship between commitment and marital quality (Surra and Gray, 2000). This dissertation investigates the relationships between commitment, marital disruption and instability, and marital quality, utilizing the context of job loss. The study examines the loss of husbands jobs at a corporation in the South around the year 2000. The job loss context is utilized, because past studies indicate that couples who have experienced job loss are more prone to marital instability and marital disruption than those who have not experienced job loss. For the purposes of research, it was important to find variation in commitment, marital stability, and marital quality. I suggest that commitment, through work at positive conditions of marriage, encourages marital quality. Findings indicate that most marriages in my sample remained successful in job loss, and all marriages remained stable. Most couples were highly committed, and small variations in commitment were not central to less success or marital disruption. Effective communication, quality time together, and positive affect seemed more important to marital quality over time than commitment; and lower commitment may be a result of decreased marital quality. Thus, commitments influence upon marital quality seems minimal, in comparison to effective communication, quality time together, and positive affect.

This project examines the relationship between patterns of crime and the development of the oil and gas extraction industry in the coastal regions of Louisiana. The suggestion of a link between these phenomena has often been made, but little systematic research has been conducted to determine if there is indeed a crime-oil development nexus. Limited previous research has focused primarily on the issue of boom and bust cycles on some forms of deviant behavior, but the data and the methods used were inadequate, and thus, the resulting conclusions were often misleading or possibly erroneous. During the course of this project, a comprehensive database is constructed that facilitates a longitudinal analysis of concomitant variation in crime patterns and oil extraction activity. Annual crime data are obtained at the parish and county level for all years beginning in 1974 and merged with corresponding social and economic data. This dataset allows for a multivariate pooled time series analysis, with adequate controls, to determine the degree of influence between oil activity and crime patterns. The results from the analysis suggest that changes in oil activity and high levels of labor market involvement in the offshore oil industry are not strongly associated with community disruption in the form of crime. The only statistically significant effects due to changes in oil activity are decreased levels of homicide and aggravated assault. Oil development is not associated with any other crime in the analysis despite accounting for the boom and bust cycles of the oil industry over a 25 year period for 12 parishes that are highly involved in the industry. As the industry becomes more active and undergoes an increased labor demand, incidents of homicide and assault decline in the community. This finding does not support some previous boomtown model research that argues that energy development causes higher rates of social disruption, including higher crime rates (Seydlitz et al. 1993a; Brookshire and DArge 1980; Dixon 1978; Finsterbusch 1982; Freudenburg and Jones 1991; Gramling and Brabant 1986).

This research attempts to shed light on the simultaneous influence of scientifically strong countries on Internet use and knowledge production at the global periphery. Using survey data from interviews of 312 Filipino scientists, this study answers the following questions: (1) Does place of graduate education (i.e., Australia, Japan, the United States and the Philippines) configure scientists Internet use? (2) Does Internet use shape scientists professional network? (3) Does place of graduate education, Internet use and professional network influence collaboration and research productivity? and (4) How does collaboration relate to productivity when professional network is accounted for? Results show that digital inequality occurs at advanced levels of hardware-software-user interaction skills, which appear to be emerging dimensions of a new form of digital inequality; these are mainly configured by level and place of graduate education. The effect of place of graduate education on networks is such that foreign training tends to increase the proportion of contacts at the scientific core. Much of the effect of the Internet lies on those components of professional network that has to do with network size, proportion of male alters, proportion of alters who are at the scientific core, location diversity, and multiplexity of communication means. Results further suggest that most scientists are involved in domestic collaboration, and that network size is positively associated with the number of collaborative projects. Whether or not networks are comprised of foreign contacts, or whether they possess a more gender-balanced configuration does not influence collaborative patterns. As regards productivity, results indicate that after relevant scientist attributes are controlled for, collaboration does not influence scientific output. There are clear indications that having a doctoral degree, possessing advanced hardware-software-user interaction skills, large networks, having more contacts at the scientific core, and proportion male alters strongly influence productivity. While professional networks influence collaboration, collaboration does not affect productivity. It could be that involvement in collaborations generates problems that undermine productivity so that scientists simply informally and causally activate their network without formally and officially engaging them in projects. It appears that this strategy is less problematic than engaging in formal collaborations.

Previous sociological research on natural disasters has highlighted how various dimensions of social vulnerability influence the impact of, and recovery from, such disasters. This research contributes to the literature by examining population change in the wake of Hurricanes Katrina and Rita, with an explicit focus on how social vulnerability moderates this relationship. Using data from the U.S. Census Bureau, I construct a macro-level Social Vulnerability Index (SoVI) for the impacted region and then use regression analysis to explore how various dimensions of social vulnerability are related to population change in the six months following the storms. The results reveal a number of significant relationships, including a history of population flux and the presence of elderly populations. However, the results are just as notable for what they do not show. Overall, I find little evidence that social vulnerability plays a major role in moderating the macro-level relationship between a disaster and population change. Implications for future research and public policy are then discussed.

This paper further develops a law-centered political process model of social movements by analyzing historical changes in American immigration law and the collective behavior of Chinese immigrants and Chinese Americans during the Chinese Exclusion Era. I present an interactive political process framework that considers not only how the broader political environment enables and constrains a movement, but also how challengers respond by actively reshaping the environment. I revisit the political process models core concept of political opportunity structure by examining legal rules and institutions, generally, and the indeterminacy of law, specifically. I apply this framework to the Chinese communitys initial use of litigation to fight the exclusion laws and their move towards direct action techniques to exploit the ambiguities of the exclusion laws. In this respect, Chinese immigrants mobilized and created legal resources.

With a rapidly aging population, maximizing independent living among the elderly is a growing concern. The inability to perform normal basic care activities interferes with an older individual's ability to live independently. In this study, I examine the onset of disability among elderly Americans using the 2002 and 2004 waves of the Health and Retirement Survey (HRS). In examining the explanatory power of both the life course perspective and the medical model, I find that the medical model is better able to explain onset of disability for males while the life course perspective prevails when explaining onset of disability among females. I find little support that living arrangements among the elderly have an impact on the probability of experiencing onset of disability. Finally, I find that differences exist in the precursors of individual activities of daily living (ADLs) disability, which suggests that using an aggregate measure of ADL disability may be masking more effective preventive measures and treatments.