This thesis studies the 1-center problem on road networks, an important problem in GIS. Using Euclidean embeddings, and reduction to fast nearest neighbor search, an approximation algorithm is devised for this problem. On real world data sets, extensive experiments are conducted that indicate fast computation of constant factor approximate solutions for query sets much larger than previously computable using exact techniques. A web interface was provided for generic queries on a map interface. / A Thesis submitted to the Department of Computer Science in partial fulfillment of the requirements for the degree of Master of Science. / Spring Semester, 2012. / March 29, 2012. / Includes bibliographical references. / Piyush Kumar, Professor Directing Thesis; Ashok Srinivasan, Committee Member; Xin Yuan, Committee Member.

A two-group randomized experimental study was used to evaluate the impact of sexual negotiation training as compared to standard health education on improving sexual risk outcomes (i.e., condom use, knowledge about a partner's sexual history, and condom use self-efficacy) among college students (n = 183). A repeated measure ANOVA revealed that, compared to those in the comparison group, participants who received sexual negotiation training were significantly more likely to report greater condom use at post-test. Secondly, intentions did not mediate the link between condition and post-test sexual risk outcomes (i.e., condom use and knowledge about a partner's sexual history). Finally, regression analyses examined whether relationship factors (i.e., relationships satisfaction, trust, commitment, communication) predicted sexual risk outcomes among participants in exclusive dating relationships (n = 108). Relationship satisfaction was significantly associated with condom use at post-test among those in exclusive relationships. Overall, the public health implications of these findings warrant attention. Findings from this study illustrate the benefits of sexual negotiation education at improving condom use and call for the development and implementation of sexual negotiation training among college students, an underserved population. / A Dissertation submitted to the Department of Family and Child Sciences in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Summer Semester, 2012. / May 29, 2012. / college students, condom use, efficacy, sexual history, sexual negotiation / Includes bibliographical references. / Frank Fincham, Professor Directing Dissertation; Mary Gerend, University Representative; Lenore McWey, Committee Member; Ming Cui, Committee Member.

The primary goal of this dissertation was to explore the use of stable isotope analysis as a tool to understand how energy flows at several layers of biological organization from the individual to the ecosystem level. With the exception of phosphorus, the primary elements that are utilized by living organisms on earth (H, C, N, O, & S) have at least one stable isotope. The isotopes of carbon, nitrogen, and sulfur are particularly useful for following the flow of energy within and between organisms, because these elements are the primary components in proteins, carbohydrates, and fats. When used in combination and when significant contrast in isotope values exists, stable isotopes can be used to elucidate biological processes from the intercellular to global scale. Chapter one will provide an introduction to the fundamentals of stable isotope analysis and a review of the state of its use in biological systems. In chapter two, the usefulness of stable isotope to identify diet changes in wild gag grouper is determined with a repeated-measures diet-shift experiment on four adult gag held in the laboratory. Fish were initially fed a diet of Atlantic mackerel, Scomber scombrus, (mean 13C = -21.3 / ± 0.2, n= 20) for a period of 56 days and then shifted to a diet of pinfish, Lagodon rhomboids, (mean 13C = -16.6 / ± 0.6, n= 20) for the 256 day experiment. I developed a non-lethal surgical procedure to obtain biopsies of the muscle, liver, and gonad tissue monthly from the same four fish. I then determined the 13C value of each tissue by isotope-ratio mass spectrometry. For the gonad tissue, I used the relationship between C/N and lipid content to correct for the influence of lipids on  13C value. I observed a significant shift in the 13C values of all of the tissues sampled in the study. Carbon turnover rates varied among the three tissues, but the shift in diet from mackerel to pinfish was clearly traceable through analysis of  13C values. The turnover rates for muscle tissue were 0.005 / day-1, and for gonad tissue was 0.009 / day-1. Although it is generally thought that tissue turnover rates in ectotherms are driven primarily by growth, I found that metabolic rate can be a major factor driving tissue turnover in adult gag. Chapter three explores the application of stable isotope analysis to trace the transfer of secondary production from temperate, coastal seagrass habitat to offshore reefs. Seagrass meadows are among the most productive ecosystems in the marine environment. It has been speculated that much of this production is exported to adjacent ecosystems via the movements of organisms. My study utilized stable isotopes to track seagrass derived production into offshore food webs in the northeastern Gulf of Mexico. I found that gag grouper (Myctereoperca microlepis) on reefs as far as 90 km from the seagrass beds incorporate a significant portion of seagrass derived biomass. The muscle tissue of gag grouper, a major fisheries species, was composed on average of 19% ±10% seagrass-habitat-derived biomass. The timing of this annual seagrass subsidy appears to be important in fueling gag grouper egg production. The 34S values of gag grouper gonad tissues varied seasonally and were34S depleted during the spawning season, indicating that gag utilize the seagrass derived biomass to support reproduction. If such large-scale trophic subsidies are typical of temperate seagrass systems, then loss of seagrass production or habitat would result in a direct loss of offshore fisheries productivity. After showing that a significant quantity of secondary production is transferred across ecosystem boundaries in chapter four, this flux is placed in the context of the flux of organic nitrogen, the primary limiting nutrient in the Gulf of Mexico. The migration of a single species, pinfish (Lagodon rhomboids), exports the equivalent of 2.85 x 107 ± 9.32 x 106 kg of primary nitrogen to the northeastern Gulf of Mexico. This flux is greater than the amount of primary nitrogen fluxed by the Apalachicola River and slightly less than the amount of organic nitrogen fluxed via atmospheric deposition and fixation by Trichodesmium. Although fish flux is less than 3 % of the total quantity of material fluxed to the offshore environment by mass, gag grouper biomass contains 19 ± 10 % seagrass derived-material. Stable carbon isotope values of otoliths can provide long-term records of the environmental conditions experienced by fish. However, as shown in chapter five because otoliths contain carbon from metabolic sources, it is necessary to determine the proportion of metabolically derived carbon to properly interpret otolith 13C values. To evaluate the relative influence of dietary carbon and ambient seawater dissolved inorganic carbon (DIC) on the 13C isotopic composition of fish otoliths, juvenile red drum (Sciaenops ocellatus) were raised for 6 and 9 months in tanks with flow-through ambient seawater and fed diets differing by 2.1 / (p < 0.001) in carbon isotope composition (13C). At the end of the experiment, muscle tissue from the two treatments reflected the isotopic composition of the diets and differed by 2.1 / (p <0.001). Muscle tissue from both groups was enriched by 1.5 / in 13C relative to the respective diets. The carbon isotope values of otoliths were enriched ~16-17 / relative to the respective diets, indicating that DIC was the dominant source (85%-92%) for otolith carbon relative to carbon derived from diet. However, otoliths from the two treatments differed by 1.6 / and expressed 60% of the difference in the isotope values of the diets. This result indicates that the nutritional characteristics of food may be a factor influencing the incorporation of metabolic carbon in the otolith. This is one of the first studies to provide evidence that the nutritional value of food influences otolith 13C value. / A Dissertation submitted to the Department of Earth, Ocean, and Atmospheric Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Fall Semester, 2011. / September 2, 2011. / ecology, ecosystem connectivity, fishereis, gag, stable isotopes, trophic transfer / Includes bibliographical references. / Jeffery Chanton, Professor Directing Dissertation; Thomas Miller, University Representative; William Landing, Committee Member; David Thistle, Committee Member; Felicia Coleman, Committee Member.

Third party presidential candidates often attract a great deal of attention, but there are multitudes of non-major party candidates at lower levels seeking to influence American politics. The following analysis will examine the contextual and individual-level causes of third party/independent voting in U. S. House and Senate elections. It will assess as well the relative success or failure of non-major party candidates at the district or state level. Chapter 2 will examine the selection of non-major party candidates by individual voters, independent of context via logit models. The influence of voters' evaluations of major party candidates on whether they select a third party/independent U. S. House or Senate candidate will be of interest. The models specify as well other demographic and attitudinal indicators, including the intensity of a respondent's partisanship, to assess their impact on vote choice. Models of vote choice in presidential elections developed by Rosenstone, Behr, and Lazarus (1996) influenced the choice of variables in these models. Their theoretical argument, as well as the one used in the following analysis, employs Hirschmann's (1970) Exit, Voice and Loyalty framework. Voters are loyal to the two-party system if they select a major party candidate, they are able to voice their concerns within that system, and they exit the two-party system if they choose a non-major party candidate. This framework will help examine the behavior of voters and electorates throughout the analysis. The behavior of electorates will be the subject of Chapter 3, where the models will determine which districts or states allocate more votes to third party/independent candidates in U. S. House and Senate elections. These models will also include the recent success of third party candidates for the office in question and the percentage of the presidential vote in the current election allocated to third party/independent candidates in the district or state. This inclusion will facilitate the examination of whether some jurisdictions are more likely to vote for third party candidates because they are less aligned with the parties than are others. Some districts may be more likely to select these kinds of candidates because their overall partisan attachments are not as strong as those of other districts. The models will also include the percentage difference between the vote shares of the major party candidates in the current contest for the office. In accordance with Burden (2007), less third party/independent voting is expected to occur in closer elections. The causes of the availability of third party/independent candidates will also be examined via Heckman (1979) selection models which will include measures of the signature requirements for ballot access and previous vote percentages of third party candidates for the office in question and the presidency. The inclusion of these percentages assesses whether district demographics show possible candidates that third party candidates can get large percentages of the vote in the state/district. These models will include the demographics of the districts to measure whether large concentrations of groups associated with the New Deal alignment, such as senior citizens and union members, deter non-major party candidates from running. The multi-level analysis in Chapter 4 will examine individual-level and macro-level factors that make voters more likely to select third party candidates. Logit-link models will examine these attributes and their impact on voting behavior; logit-link models can take into account variables measured at the district/state level (i.e., the percentage difference between major party vote shares) and the individual level (i.e., respondents' evaluations of major party candidates), and evaluate their influence on individual vote choice. The models will include individual-level variables from Chapter 2, as well as some macro-level indicators. The individual variables in the logit-link models will include: the last third party vote percentage for the office being contested, the current or most recent third party/independent presidential vote percentage, and the percentage difference between the major party candidates in the current election. These variables are expected to send the signal to voters that exiting (Hirschmann 1970) the two-party system is a viable choice in their area. Context is expected to condition whether a voter chooses a candidate who is not aligned with the major parties, independent of individual preferences. In the analysis, a respondent's evaluations of major party candidates are a significant predictor of his/her vote choice in all of the models that use this measure of a respondent's preferences. Also, higher current or most recent third party/independent presidential vote percentages make respondents more likely to select non-major party candidates in all of the logit-link models that use this variable to predict vote choice. Candidate preferences are very strong determinants of individual vote choice, and larger percentages of third party presidential votes in a district/state make respondents much more likely to select third party legislative candidates. / A Dissertation submitted to the Department of Political Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Fall Semester, 2011. / August 29, 2011. / Congressional Elections, Political Parties, Third Parties, Voting / Includes bibliographical references. / Robert A. Jackson, Professor Directing Dissertation; Lance deHaven Smith, University Representative; Charles Barrilleaux, Committee Member; Jennifer Jerit, Committee Member.

This dissertation consists of three essays. The first two essays (i.e., Chapter 2 and Chapter 3) examine the effects of raising the retirement age on the life cycle behaviors of individuals and its implication on the social security budget. The third essay (i.e., Chapter 4) is an empirical study, which tests the hypothesis of investors' overreactions when trading neglected stocks. The first essay examines the impact of raising the retirement age on the saving and working behaviors of older individuals, and the associated impact on the social security budget. Its results indicate that the reform would result in a 50% reduction in the social security budget deficit. In terms of behavioral responses, we find that: (1) individuals respond to the reform by saving more progressively during the period prior to retirement (i.e., from their early 40s to age 62), while supplying more working hours during the retirement period (i.e., ages 62 and older). The intensity of the saving and working hour responses critically depend on the assumption of the efficiency indexes of the elderly: the lower (higher) the efficiency index, the more intense the saving (working hours) response; (2) there is an upward shift in the working hour profile of individuals as a result of raising the retirement age. Once again, the distance of the shift increases with values of the elderly efficiency index; (3) we find a decrease in the participation rate of elderly individuals age 62-80 in versions where the estimated efficiency index of the elderly is relatively low. The second essay focuses on examining the life-cycle behavior responses of individuals with different skill levels to the raising of the retirement age reform. We find that individuals with different educational attainment respond differently to the reform. Specifically, individuals with lower-than-average education respond to the policy change with a significant upward shift in the working hour profile, a higher participation rate, and an aggressive retirement saving motive. On the other hand, individuals with higher-than-average education mainly deal with the policy change by a higher saving rate and/or a lower rate of decumulating their assets in the retirement period. More importantly, the participation rate in the retirement period among these individuals is actually lower than before the policy change. Secondly, our findings suggest that disadvantaged individuals (e.g., those with a low education level) are the ones who are heavily affected by the policy reform in terms of a bigger consumption reduction, a more intense labor supply response, and a higher contribution to the social security budget. Finally, we find a small increase in the average labor productivity associated with the policy change. However, by educational attainment, we find evidence which suggests a decrease in labor productivity among individuals with below-average educational attainment (i.e., those with a high school degree or lesser), and an increase in labor productivity among those with above-average educational attainment (i.e., those with a college degree or higher). The third essay is an empirical study, which tests the hypothesis of investors' overreaction when trading stocks with limited information, such as neglected stocks. Specifically, we design a fundamental scoring method (called NSCORE) and apply it to the neglected stock universe. We also apply this method to the most-watched stock universe (called WSCORE). Our results show that the annualized returns of a monthly-rebalancing investment strategy which buys the top 100 NSCORE and sells the bottom 100 NSCORE is 26.31% for the period from the beginning of 1985 to the end of 2009. By contrast, when applying the same screening method to the most-watched stocks universe during the same time period, the annualized returns of the same investment strategy dropped to about half. This evidence clearly demonstrates the effectiveness of using financial statement data to identify winners and losers among neglected stocks as a result of investors' overreaction. We also find that the returns difference between top and bottom neglected stocks tends to persist for a long time. Specifically, the return difference between the top 100 NSCORE and the bottom 100 NSCORE can last up to 36 months (3 years). On the other hand, the returns difference among most watched-stocks tends to generally disappear after 12 months (1 year). Our comprehensive sensitivity tests confirm that our findings are not subject to well-known anomalies such as the size, book-to-market, and illiquidity effects. / A Dissertation submitted to the Department of Economics in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Spring Semester, 2012. / January 3, 2012. / Life Cycle Labor Supply, Neglected Stocks, Social Security Reform / Includes bibliographical references. / Don Schlagenhaulf, Professor Directing Dissertation; Alec N. Kercheval, University Representative; James Cobbe, Committee Member; Paul Beaumont, Committee Member.

ABSTRACT The purpose of this study is to develop a better understanding of the Freedom of Information Act (FOIA) policy formulation and implementation by analyzing the Bill Clinton and George W. Bush administrations' FOIA policies. The problem this study addresses is why the Clinton and Bush administrations pursued different FOIA policies even though it appears that "an informed citizenry" was a basic FOIA principle shared by federal FOIA employees through both administrations. This study assumes that the President's comments and statements greatly affect the actions and decisions of the Executive Branch. This study used the principal agent theory, which identifies "hierarchical control," "goal conflict" and "difficulty in monitoring" as significant concepts. To answer the research questions, this study employed multi-qualitative methods, which are mainly non-reactive or unobtrusive research methods including content analysis, secondary analysis and document analysis. The author collected quantitative data from the OIP newsletter, the FOIA Update (1993 to 2000) and the FOIA Post (2001 to 2006), distributed quarterly in paper format until 2000. The FOIA was not a main agenda item of the Clinton and Bush administrations, although both Presidents Clinton and Bush showed some interest in the FOIA. The president's role in FOIA policy formulation is more than symbolic; Presidents Clinton and Bush had different political philosophies regarding the FOIA. Clinton considered the FOIA an essential facet of democracy, whereas Bush considered that the FOIA could be limited for national security, effectiveness of government performance, and personal privacy; The September 11, 2001 terrorist attacks seemed to have added impetus to extend the Bush administration's restrictive FOIA policy, accelerating the administration's drive to regain presidential power. The Ashcroft memorandum and the Card memorandum seemed to change the climate of FOIA implementation from encouraging information release to protecting national security information. The similarities in FOIA policy between the two administrations are issuing FOIA directives, user-friendly ways, and acknowledgement of the importance of national security, effective government performance, and privacy. The differences in FOIA policy between the two administrations are opposite FOIA initiatives, different political environments, and structural changes for FOIA organization. Finally, the main implications of this study are that the president's philosophy on the FOIA had effects on federal FOIA policies; high level officers and political appointees were also able to affect FOIA policy formulation and implementation; middle-level FOIA officers had a critical role in FOIA implementation, a dual role in which they served both as principal and as agent; FOIA culture seemed to affect federal departments' FOIA implementation; insufficient and poor guidance have been a major hindrance to FOIA implementation; and Congress is one of the most important principals in FOIA policy formulation. / A Dissertation submitted to the School of Library and Information Studies in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Fall Semester, 2011. / October 17, 2011. / Bill Clinton, FOIA, George W. Bush, information policy, principal agent theory / Includes bibliographical references. / Charles McClure, Professor Directing Dissertation; Gary Burnett, Committee Member; Chris Hinnant, Committee Member.

The study explored instructional benefits of a storyline gaming characteristic (GC) on learning effectiveness, efficiency, and engagement with the use of an online instructional simulation for graduate students in an introductory statistics course. In addition, the study focused on examining the effects of a storyline GC on specific learning outcomes, i.e., factual, conceptual, and application knowledge. In order to study the interactions between the storyline GC and human performance, a storyline was embedded in an instructional simulation. The goal of the simulation was to engage students in problem solving and data analysis in the context of basic statistics and allow students to practice the newly acquired skills for analyzing real-world examples. The author developed two different versions of the instructional simulation: (1) Simulation+No GC and (2) Simulation+Storyline GC. All versions had the same instructional content but differed in presence or absence of a storyline GC. Sixty-four graduate students with basic statistical knowledge participated in the study. Students were randomly assigned to two intervention conditions: (1) Simulation+No GC, and (2) Simulation+Storyline GC. Each of the intervention groups completed an online instructional module that required approximately two hours for a learner without a substantial background in the subject matter. During the instructional module, participants were engaged in: (1) an online instruction presenting statistics concepts to-be-learned, (2) a pretest assessing students' prerequisite and prior knowledge of the subject matter, (3) a simulation-based practice, (4) a posttest assessing the taught skills and concepts, and (5) demographic and engagement surveys. The results indicated that adding a storyline to a learning simulation did not result in significant improvements in learning effectiveness or efficiency. The analysis of students' performance scores for factual, conceptual and application knowledge did not reveal any significant differences between the interventions as well. However, students' performance from both interventions combined significantly improved from pre- to posttest. Both interventions (simulation and simulation with a storyline) showed significant learning gains related to application knowledge, thus supporting previous research showing a positive effect of simulations and games on developing higher cognitive skills. With regard to engagement, significantly higher engagement levels were observed among the students from the Simulation+No GC than the Simulation+Storyline GC group. The findings of this study suggest that adding a storyline may distract students from an instructional task and pose a higher level of extraneous cognitive load that might be detrimental to learning outcomes. Moreover, students' preferences as well as the nature of the storyline and the storytelling system can affect students' engagement in the learning process. Designing games with adaptive instructional gaming characteristics that reflect students' preferences may improve students' performance and engagement. / A Dissertation submitted to the Department of Educational Psychology and Learning Systems in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Spring Semester, 2012. / November 14, 2011. / games, gaming characteristics, simulations, statistics education, storyline / Includes bibliographical references. / Tristan E. Johnson, Professor Directing Dissertation; Patrice Iatarola, University Representative; Valerie Shute, Committee Member; Gershon Tenenbaum, Committee Member.

This work analyzes the theme of the Eternal Feminine and its significance within Goethe`s famous drama Faust. The first chapter delineates and explores the various rich definitions of the Eternal Feminine found in the realms of theology, philosophy, and literature. The chapter concludes with the assertion that although the definitions of the Eternal Feminine are myriad and complex and that there is no one clear definition, that a careful inspection of these definitions allows one to ascribe to the concept of the Eternal Feminine certain characteristics. Thus, the Eternal Feminine is Beauty, Truth, Good, Love. The second chapter addresses in detail the theme of evil as represented in Faust with the purpose to show that the Goethe`s Eternal Feminine is a form of salvation and therefore cannot be understood without a discussion of the theme of evil to which it acts as an antidote. The chapter proceeds with the analysis of the role of the Eternal Masculine or the Father Figure as represented in Faust and its relation to the Eternal Feminine in the drama. The third chapter focuses on the various representations of the Eternal Feminine in Faust (as the Perfect World of Platonic Ideals (Eros), as the Indian Feminine Goddess, the Fruitful Mother Nature, and the Sublimated Sexual Energy). Finally, the Platonic and Pagan elements in Goethe`s idea of the Eternal Feminine are examined and contrasted with the Christian idea of salvation. A conclusion is asserted that Goethe`s Salvation via Eternal Feminine is the return to the pre-Christian search for life`s meaning and redemption from suffering and evil. / A Thesis submitted to the Department of Modern Languages and Linguistics in partial fulfillment of the requirements for the degree of Master of Arts. / Spring Semester, 2012. / March 15, 2012. / Eternal, Faust, Feminine, Goethe / Includes bibliographical references. / Christian Weber, Professor Directing Thesis; Birgit Maier-Katkin, Committee Member; Reinier Leushuis, Committee Member.

The first essay conducts robustness analysis on Gali's (1999) results. Following Gali's identification strategy, the model is extended to the sectoral level within the private sector. The paper also looks at the two important breaks, 1973 recession and 1984-beginning of the "great moderation". The private sector results suggest that non-technology shocks are the major cause of business cycle fluctuation rather than technology shocks. Sectoral data also produced this conclusion with the exception of one sector. Most of the results do not change for the pre- and post-recession and great moderation dates. This essay reinforces the notion that technology shocks play a limited role in the aggregate short-run fluctuations of business cycles. These results pose a challenge to modern real business cycle theory. The question does hours decline in response to a technology shock attracted a lot of research in the last decade. The second essay attempted to investigate the response to hours in a three-variable--productivity, hours and corporate profits-- model using vector- autoregressive with long-run and short-run restrictions. The model imposes three restric- tions: technology shocks affect productivity permanently, hour's shock and profit shocks do not affect productivity in the long-run and profit shocks do not affect hours contempora- neously. The results seemed to be more encouraging for real business cycle theory and are inconsistent with the conclusion that technology shocks play limited role in business cycle fluctuations. An important finding is that profits matter empirically since it changed the response to hours from a technology shock. By adding profits to the model, hours do not decline from a productivity shock. Though the initial impact is negative they recover in first quarter and they co-move with productivity.The response to hours shock is however consistent with Gali (1999). Hours worked increase in response to a shock to employment. Recent empirical research argued that intangible capital has been playing an important role in explaining productivity gains in the last two decades. In the third essay, intangible capital is introduced in an otherwise standard real business cycle model. Firms expend resources to create intangible capital which is an additional input in the production func- tion. Since firm's investment in intangible capital is pro-cyclical it produces positive profits despite being a competitive firm. The firm increases investment in intangible capital from both temporary and permanent productivity shock. It also plays a significant role in pro- ducing endogenous movement in productivity. Firms use more labor and physical capital to produce intangible capital since it raises productivity and future profits. However, there is a trade-off between current period profits and investment in intangible capital. Perma- nent technology shock results in higher factor share of labor and capital allocated to create intangible capital which decreases profits in the current period; however, higher investment in intangible capital would raise future profits. / A Dissertation submitted to the Department of Economics in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Fall Semester, 2011. / September 7, 2011. / Business cycles, Corporate Profits, Intangible Capital, Long-Run and Short-Run Restrictions, Productivity, Technology Shocks / Includes bibliographical references. / Milton H. Marquis, Professor Directing Dissertation; Alec N. Kercheval, University Representative; Stefan C. Norrbin, Committee Member; Paul M. Beaumont, Committee Member.

Social Anxiety Disorder (SAD) is a prevalent disorder marked by significant impairment. Advancing our understanding of the disorder is important to prevention and treatment of SAD. Formal diagnostic criteria and traditional accounts of SAD highlight the feature of avoidance; however, this depiction of SAD minimizes the full spectrum of responses to social anxiety. If humans are evolutionarily prepared to face perceived threat with a "flight or fight" response, it seems reasonable to assume that the entire nature of this response should be observed in clinical manifestations of anxiety. Following work by Kashdan and colleagues (Kashdan, Elhai, & Breen, 2008; Kashdan & Hoffman, 2008; Kashdan, McKnight, Richey, & Hoffman, 2009), this study directly examined this confrontational side of responses in social anxiety, specifically aggression, in a sample of socially anxious college students. Two hundred twenty-one students who exhibited high scores on two measures of social anxiety were subjected to a paradigm in which they were asked to make decisions ostensibly related to painful tasks and hurtful ratings of partners in the study. In order to increase the likelihood of aggressive responding, half of the participants were told that the study was investigating leadership skills, and their leadership skills were called into question. All participants received measures of symptoms/traits thought to relate to aggressive responding; a component score was derived from these measures to aid in the detection of participants most likely to aggress. Results showed that some participants did display aggressive responding in the study. Direct aggression was predicted by a three-way interaction between component score, experimental manipulation, and sex, and approximately 16-22% of the sample displayed high levels of aggression on the various tasks measuring direct aggression. Traits of paranoia, psychopathy, and borderline personality were most related to direct aggression. Indirect aggression was predicted by component score, and 21.27% of the sample displayed high levels of indirect aggression. Traits of paranoia were most related to indirect aggression. Self-reported aggression was predicted by component score and gender, and 7.78% of males in the sample and 14.50% of females in the sample reported high levels of self-reported aggression. Traits of paranoia, psychopathy, and depressive personality were most related to self-reported aggression. These patterns of results were not, however, mediated by beliefs about the status enhancement potential of aggression. The results, as a whole, support prior findings of atypical responses to social anxiety and suggest that some social anxious individuals may respond with aggression rather than prototypical avoidance. / A Dissertation submitted to the Department of Psychology in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Summer Semester, 2012. / June 26, 2012. / Aggression, Social Anxiety / Includes bibliographical references. / Norman B. Schmidt, Professor Directing Dissertation; Bruce A. Thyer, University Representative; Thomas E. Joiner, Committee Member; Jon K. Maner, Committee Member; Jesse R. Cougle, Committee Member.