This dissertation examined the performance outcomes of public management networks (PMNs) in the disaster management context. The effectiveness of three disaster response sub-networks in the area of evacuation were examined and compared using the case of the Hurricane Katrina disaster in New Orleans, Louisiana in August 2005: Citizen Protection: Evacuation and Shelter, Emergency Public Information and Warning, and Public Safety and Security Response. Although more public sector programs and services mandated by federal policy are being delivered by PMNs, the effectiveness of such networks has not been systematically evaluated. Two primary research questions guided this dissertation: "To what extent does governance influence network effectiveness?" "To what extent does the structure of a network and its internal communication processes influence network effectiveness under conditions of uncertainty?" Four theoretical perspectives were used to examine the research questions and hypotheses: institutional, social network, resource dependence, and electronic proximity. A mixed methods approach was used that consisted of a comparative case study using grounded theory analysis, social network methods, and a ratio measure of the percentage of goals accomplished by each sub-network. Quadratic Assignment Procedure (QAP) regression, QAP correlations, and linear regression were conducted to analyze the effects of network effectiveness on the independent variables. Four determinants emerged from the grounded theory analysis: political discord, administrative incompetence, lack of readiness, and insufficient critical resources. In-depth interviews with senior and mid-level public sector managers suggested that these determinants negatively influenced the effectiveness of the disaster response sub-networks during Hurricane Katrina. The task goals of the three sub-networks were not closely monitored by federal governance oversight agencies to ensure compliance with federal disaster response policies during the year prior to the Hurricane Katrina disaster. The results of the QAP regression showed that information sharing and information and communication technology (ICT) use among organizations were significant, positive predictors of network effectiveness. However, the results of the linear regression showed that the presence of centralized authorities was not significant in the three sub-networks. The level of integration in sub-networks two and three was not statistically significant. The network analysis confirmed that network ties were not formed between some organizations and ties were broken during the disaster response; information silos developed between organizations in each sub-network. Although organizations in all three sub-networks became disconnected during the disaster response, those that were able to consistently engage in information sharing and use ICTs were more likely to complete disaster response goals. Overall, the three sub-networks were moderately effective in the delivery of disaster response services, which reflected the percentage of network ties that were connected during the disaster response: approximately 51% for sub-network one, 49% for sub-network two, and 61% for sub-network three. The effectiveness of each sub-network was hindered by its level of integration. Network effectiveness was influenced by the level of capacity and inadequate resources in each sub-network that resulted in the accomplishment of 50% of the network goals for sub-networks one and three, and 67% for sub-network two. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]