Youth Tobacco Surveys have been conducted in several states in the U.S. in recent years, in order to design policies with the goal of reducing tobacco use among young people. Some primary analysis of those surveys has been done, but few analyses include modeling, and the study of independence has been addressed, mainly, in the bivariate context. In this work contemporary methods, which are of relative recent appearance in categorical data analysis, will be examined, including logistic and log-linear modeling as well as graphical displays and correspondence analysis. These methods will be applied to data from the 2000 Tennessee Youth Tobacco Survey. The objective is to demonstrate that methods of multivariate categorical data analysis can provide fresh insight about the behavior of adolescents with respect to tobacco use. The ultimate purpose of this work is to recommend methodology that goes beyond that which is currently published.

This paper investigates variation of lexical and analytic causatives in 15 European languages from the Germanic, Romance, and Slavic genera based on a multilingual parallel corpus of film subtitles. Using typological parameters of variation of causatives from the literature, this study tests which parameters are relevant for the choice between analytic and lexical causatives in the sample of languages. The main research question is whether the variation is constrained by one semantic dimension, namely, the conceptual integration of the causing and caused events, as suggested by previous research on iconicity in language, or whether several different semantic and syntactic factors are at play. To answer this question, I use an exploratory multivariate technique for categorical data (Multiple Correspondence Analysis with supplementary points) and conditional random forests, a nonparametric regression and classification method. The study demonstrates the importance of corpus data in testing typological hypotheses.

Skärvstenshögarna i Uppland har konstaterats vara en komplex fornlämningstyp och det har visat sig vara problematiskt att förklara dem som homogena i form och funktion över tid. En kategorisering av dem har skapats som har konstaterat att olika former av dem finns. Syftet med denna uppsats är sedermera att utgå från befintlig kategorisering och undersöka om fler kategorier och typer går att isolera ur tillgänglig empiri. Därefter för att beskriva deras form, läge, kronologi och förändring. Teoretiska ramverk som används för att begreppsliggöra skärvstensbrukets mer än tusenåriga användande är de om tradition, minne och ritual. Teorier som kan manifestera bruket som en kulturell process och dimension i tid och rum. Metoder som används för att identifiera fler kategorier och typer är dels multipel korrespondensanalys av 100 skärvstensförekomster för att generera mönster. Dels densitetsanalys av 100 C14-dateringar av detsamma för att utarbeta aspekter av tid i empirin. Slutligen används visibilitetsanalys för att generera belägg för kategoriernas olika lägen samt en kvalitativ genomgång av empirin för att skapa modeller av nya typer inom befintliga kategorier. / The burnt mounds of Uppland have been shown to be remains of a complex type and it has been proven problematic to explain them as homogeneous in form and function over time. A categorization of them has been created which has established that different forms exist. From that categorization the purpose of this essay is subsequently to investigate whether additional categories and types can be isolated from available empirical evidence. Also to include a description of their form, location, chronology, and change. Theoretical frameworks used to conceptualize the more than a thousand-year use of burnt stones are those of tradition, memory, and ritual. Theories that can manifest the practice as a cultural process and a dimension of time and space. Methods used to identify additional categories and types includes a multiple correspondence analysis of 100 burnt mounds to generate patterns. A density analysis of 100 C14 samples of the same deposits to work out aspects of time in the empirics. Finally, visibility analysis is used to ascertain the different positions of the categories as well as a qualitative review of the empirical evidence to create models of new types within existing categories.

A population survey was conducted from April through September 2002 on mosquito species occurring on the Ray Roberts Greenbelt, a riparian corridor used for public recreation on the Elm Fork of the Trinity River, in Denton County, Texas. ArcGIS software was used to set up a stratified random sampling design based on habitat parameters. Multivariate analyses of sampling data and climatic variables were used to describe spatial and temporal patterns of mosquito species. A total of 33 species were collected during this study belonging to the following genera: Aedes, Anopheles, Coquillettidia, Culex, Mansonia, Ochlerotatus, Orthopodomyia, Psorophora, Toxorhynchites, and Uranotaenia. Seasonal distributions of the dominant species revealed population fluctuations. Aedes vexans was the primary species collected in April and May, occurring in low numbers throughout the rest of the sampling period. Psorophora columbiae reached its highest population density in June, with a smaller peak occurring in late July. Present from May through the end of September, Culex erraticus was the most abundant species collected with major peaks in mid-June and the end of July. Abundance of Culex salinarius followed the same general trend as that for Cx. erraticus, but with smaller numbers. The specimens were tested for a variety of arboviruses by the Texas Department of Health. One pool of Cx. erraticus and Cx. salinarius, collected in August 2002, tested positive for West Nile virus. Variables that were important factors for determining dominant species abundance were temperature, wind speed, rain accumulation occurring one-week and two-weeks prior to sampling, number of day since last rain event, dew point, and average canopy coverage.

Nós investigamos as relações entre a diversidade e a estrutura do componente arbóreo com as variáveis ambientais, em quatro posições microtopográficas (MPs), ao longo de um gradiente formado entre zonas úmidas depressionais florestadas e terras firmes em jaboticabais (FDWs) no Planalto Sul-Brasileiro. A riqueza específica e a diversidade foram similares nas terras firmes e menores nas zonas úmidas, enquanto que os parâmetros estruturais diferiram pouco entre as MPs. Uma análise de espécies indicadoras mostrou que poucas espécies são indicadoras das terras firmes, revelando que a distribuição das espécies foi relativamente uniforme em comparação às zonas úmidas. Os mesmos padrões foram revelados por uma análise de correspondência canônica (CCA), que mostrou que os alagamentos foram os principais determinantes da zonação de espécies arbóreas. Aparentemente, Plinia peruviana (Poir.) Govaerts (Myrtaceae) é a única espécie com distribuição praticamente restrita às MPs estudadas, mas os jaboticabais podem ter uma importância desproporcionalmente maior na preservação de espécies que são pouco abundantes em outras florestas na região. A ocorrência exclusiva de solos distróficos nas áreas de estudo, em uma região onde solos eutróficos predominam, indicam que as terras firmes estudadas representam zonas transicionais entre as zonas úmidas e a matriz florestal. Como estudo pioneiro em zonas úmidas depressionais florestadas no Planalto Sul-Brasileiro, nossos resultados sugerem que ao menos toda a área correspondente às depressões topográficas deve ser protegida, até que estudos complementares proponham zonas de amortecimento apropriadas, com base em critérios biológicos. Jaboticabais são ecossistemas únicos cuja restauração, manejo e preservação são cruciais para manter ambientes espacialmente heterogêneos e, consequentemente, a riqueza específica e as funções e serviços ecossistêmicos. / We investigated the relationships between tree component diversity and structure with environmental variables of four microtopographic positions (MPs) along a wetland-upland gradient in forested depressional wetlands (FDWs) on the South Brazilian Plateau. Species richness and diversity were quite similar in the uplands and lowest in the wetlands, while structural parameters differed little among MPs. An indicator species analysis showed that few species were indicators of the upland sites, suggesting that species distribution was relatively uniform in comparison to the wetlands. The same patterns were revealed by a canonical correspondence analysis (CCA) that showed flooding to be the main factor driving tree species zonation. Apparently, Plinia peruviana (Poir.) Govaerts (Myrtaceae) is the only tree species with practically restricted distribution to the studied ecosystems, although FDWs and their surrounding uplands can have a disproportionately higher importance in the preservation of species that are under-represented in other regional forests. The occurrence of exclusively dystrophic soils in the study areas, in a region where eutrophic soils predominate, indicates that the studied uplands represent transitional zones between the FDWs and the forest matrix. As a pioneer study in FDWs on the South Brazilian Plateau, our findings suggest that the whole area of the topographic depressions should be protected at least until complementary studies propose proper buffer zones based on biological criteria. FDWs are unique ecosystems whose restoration, management and preservation are crucial to maintain spatial environmental heterogeneity and consequently species richness and ecosystem functions and services.

Savanna woodlands are vitally important in providing ecological services (e.g. erosion protection, micro-climate) and economic services (e.g. timber, food, fodder non-wood products, and wild-life habitats) that sustain local livelihoods and national economies. Increasing demands and the need for sustainable savanna woodland resource management requires that the ecological, economic, social and cultural values of these resources be explored and brought to the attention of decision makers and the general public. The identification and better understanding of the structure and dynamics of woodland community types, patterns of species distribution and quantitative properties of their diversity is important to the conservation and sustainable management of these woodlands. This study seeks to contribute to a better understanding of Nakasongola woodland community types, species diversity patterns and environment correlates, natural regeneration processes (i.e. sprouting and seedling establishment) and identifying livelihood strategies adopted by households, woody species utilised, and the contribution of charcoal production to household livelihoods. Data on vegetation and environmental variables were collected using 75 rectangular 20 x 50 m (0.1 ha) plots. Data on land use and land cover changes, and relevant associated socio-economic parameters were collected through the analysis of multi-temporal satellite imagery and field observations, as well as interviews of local households and key informants. The basic major livelihood activities for the rural households in this savanna dryland are charcoal production, subsistence crop cultivation and livestock grazing. However, it, sometimes, includes various combinations of activities, i.e. charcoal production and subsistence crop cultivation for both food and cash, and livestock keeping for income generation through selling the livestock products such as milk and, sometimes, the whole animal. At least 24 woody species, including fruit trees (Mangifera indica and Artocarpus heterophyllus), are frequently harvested, including 16 species that are considered the most utilized for charcoal production. Charcoal production, being the major source of income to the rural households, contributes on average US$ 259 ± 46 (S.E.) per household annually. There were significant differences in charcoal production (Kruskal-Wallis; H = 31.42, p < 0.0001), producer sale prices per bag of charcoal (H = 35.62, p < 0.0001), and annual incomes from charcoal production (H = 32.44, p < 0.0001) per households across the 8 sub-counties. Most of the youth (≤ 20 years old) derive their livelihoods from charcoal production, a small amount of trade, offering labour services, livestock keeping, fishing, bee keeping and earth brick making. Charcoal production, livestock keeping and hunting are carried out particularly by men, whereas, crop cultivation, and collection of fire wood, medicinal plants and fruits are carried out, mainly, by women. However, men are also engaged in cultivation only during the rainy seasons. There have been significant land cover changes in the area during the period 1984 to 2001, resulting in a 64% decrease in dense woodland cover, and an 80% increase in areas under cultivation/settlements. These changes are attributed to significant spatial expansion in agriculture increased commercialisation of charcoal production, grazing and human population growth. A total of 44,195 (5,893 plants/ha) woody plants representing 99 species in 67 genera and 31 families were recorded. The most species rich families were Mimosaceae (13), Rubiaceae (9), Moraceae (7), Euphorbiaceae (7), Anacardiaceae (6), Combretaceae (5) and Verbenaceae (5). Density of woody species differed significantly (F2, 72 = 6.3, P < 0.003) among land uses, being higher under charcoal production (7,131 ± 755 plants/ha) and cultivation (6,612 ± 665) areas and significantly lower under grazing lands (4,152 ± 525). Community species composition differed significantly (Global RANOSIM = 0.14, p = 0.001) among land use types. All measures of beta-diversity (spatial “turnover” in species composition) showed consistently higher beta-diversity in the grazing land use (βW = 3.1; βT = 3.1), followed by cultivation (βW = 2.8; βT = 3.0) and charcoal production (βW = 2.7; βT = 2.8), suggesting a more heterogeneous spatial distribution of species in the grazing lands. This suggests that variations in the composition and diversity of woody species are to a great extent influenced by land use type and anthropogenic disturbances in this region. Basal area of woody species differed significantly (F2, 72 = 12.0, P < 0.0001) among land uses, being highest under cultivation and charcoal production and significantly lower under grazing. Woody plant density differed (F2,72 = 6.3, P = 0.003) across landuses, being highest under charcoal production and cultivation and significantly lower under grazing. The species that contributed most to both basal area and density across all the land uses were Combretum collinum and Combretum molle. However, different species contributed the next most i.e. Piliostigma thonningii for grazing; Albizia zygia and Harrisonia abyssinica for cultivation and Vepris nobilis for charcoal production areas. For both basal area and abundance of all woody species, the total variance in species-environmental factor relations (for the combined first four canonical axes) was higher than 50%, suggesting a relatively strong influence of the measured environment variables on species composition and distributions. The CCA points to a significant influence of soil Ca2+ and Mg2+ in association with grazing on gradients in the composition and structure of woody species in the savanna woodland of Nakasongola. Resprouting was generally common among the woody species. A total of 2,595 stumps, representing 74 species in 31 families were recorded from all plots. Of these, 98.3% resprouted and were identified to species level. Density of both stumps and total resprout differed significantly (p < 0.05) among the land uses, being higher in charcoal production areas than in grazing and cultivation land uses. For the overall pooled data, resprouts per stump differed significantly among land uses (F2, 456 = 7.75, p = 0.0005), being highest in charcoal production (mean ± S.E.; 14 ± 1) and cultivation (13 ± 1) land uses and lowest under grazing areas (10 ± 1). Generally, the mean number of resprouts per stump increased with increasing stump basal diameter (BD), being highest for BD size class > 41 cm. In relation to stump height, the highest mean resprouts/stump was found on stumps with heights ranging from 0.31- 0.40 m. Based on pooled species data, regression analyses showed weakly significant negative relationships between BD of leading resprouts and number of resprouts/stump (r2 = 0.123, p < 0.0001) and between height of leading resprouts and number of resprouts/stump (r2 = 0.068, p < 0.0001). Density of seedlings of woody species differed significantly among land use types (ANOVA; F2, 72 =5.9, p = 0.004), being highest for cultivation (3,162 ± 440 individuals ha-1), followed by charcoal production (2,416 ± 295 ha-1) and lowest for grazing (1,629 ± 205 ha-1). Composition of seedlings differed significantly among land use types (Global RANOSIM = 0.119, p = 0.001). The distributions and densities of some seedlings were explained by gradients in environmental variables, with edaphic factors (i.e. Ca2+, Mg2+, K+ and organic matter) and charcoal production being the most important. The first two axes of the Canonical Correspondence Analysis (CCA) explained 41.9% of the variance in species – environmental relations and were a reflection of edaphic and charcoal production land use gradients. All of the 16 highly utilized species were well represented in both the juvenile and adult classes, with gradually declining number of individuals with increasing stem size-class. This indicates that most of these species have high regeneration potential. Juvenile:adult tree ratios >1 and negative DSCD slopes indicate good recruitment and probably successful regeneration for these species. The study revealed land cover changes mainly in the dense and medium dense woodlands, reflected by the increase in open woodland, grasslands and cultivation/settlements. These trends threaten the livelihoods of local communities who are entirely dependent on these natural resources. Sustainable management will require the establishment of suitable integrated community-based institutions and management practices, with support from all key stakeholders (i.e. National Forest Authority (NFA)) and local communities. Maintenance of savanna woodland resources and other ecosystem services essential for human well-being will require an effective legal framework to prevent over-exploitation and give incentives for the protection of the fragile savanna woodland vegetation. An appropriate savanna woodland management policy will be required to guide changes in land use that accommodate the requirements of land users, aided by targeted conservation efforts to all woody plants and particularly for the highly utilized species for charcoal production as well as the multipurpose species. In addition, there is urgent need to build local capacity for improved harvesting and utilization of these tree species. This can be achieved through equipping local users with up to date information as well as observing the existing skills.

Figurines with a human shape have been created in almost every culture all throughout human history. In this thesis one such group of figurines is under investigation, Viking age female figurines. These figurines are most often interpreted as representations of the Valkyries – shieldmaidens of the god Odin – or as the goddess Fröja. Interpretations made through comparative studies with the medieval written sources. However, these interpretations always privileges some attributes of the figurines over others, creating simplified and general interpretations. The purpose of this thesis is to (re)interpret the figurines beyond the concept of representation, and instead focus on what the figurines, through their various attributes (size, motive, material), do and how they influence human actions. This is realized through the creation of a catalogue of all currently known female figurines from the Viking age (53 pieces), a correspondence analysis and through the use of symmetrical archaeology and embodiment theory. The result shows that the most important attributes of the figurines is their physical bodies, through which they can interact with the world, trigger emotions, hold memory, become animated and be part of the performative practice of upholding individual identity as well as upholding both the social- and cosmological worlds.

本文是就有序變數的模型在 個不同“維度”的架構下做探討，主要可以區分成兩部份。 1.就兩個有序變數的二維列聯表，我們介紹ＲＣ（Ｍ）模型，並分別探討關聯性模型，相關性模型及對應分析模型。 2.就三個變數的三維列聯表，在給定一個變數的情形之下，我們討論了如何運用三變數的ＲＣ（Ｍ）關聯性模型來探討另外兩個有序變數間的關係。 / We examine statistical models for ordinal variables that allow for more then one“ dimension ”of association in this study. The focus are on the following two types of models: 1. RC(M) models for two ordinal variables, including RC(M) association model; RC(M) correlation model and RC(M) correspondence analysis model. 2. Conditional RC(M) association models which can be used to analyze a 3-way contingency table with at least two ordinal variables.

傳統的犯罪學在於研究犯罪者生理與心理的特性以及外在環境因素，藉以 尋求預防犯罪的策略。其實，預防犯罪的目的是要減少被害者數目。因此 ，吾人希望藉由探討被害者家庭形態的特質與社會犯罪的關係，使防範被 害的政策能夠落實到家庭層面。所以本研究的目的就是要利用統計方法， 透過數字及圖形上的表達，來描述住戶之家計負責人的特性(如性別、年 齡、教育程度)與住戶遭受侵害的事實(曾或否)及類型(如財物侵害、暴力 侵害、其他侵害)之間的關係，期為犯罪統計工作提供一些分析的工具。 在本篇論文，首先利用對數線性模式分析，選出最適合的模式，以解釋家 計負責人之性別、年齡、教育程度與住戶是否遭受侵害等四個變數之間的 關聯性。接著利用推理值模式分析探討家計負責人之性別，年齡，教育程 度這三個解釋變數各分級對住戶遭受侵害之影響的主效應及交互作用效應 。最後則是利用對應分析處理資料，透過對應分析的顯示圖形，可以用來 解釋家計負責人之性別、年齡、教育程度以及住戶遭受侵害的犯罪類型之 間的對應關係。

Woody invasion into grasslands has been reported world-wide and has affected both the magnitude and spatial heterogeneity of soil carbon (C) and nitrogen (N). Since grasslands cover a large portion of the Earth's land surface, invasion of woody plants could have impacts on regional and global biogeochemistry. To understand large-scale ecological and policy implications of woody invasion, it is critical to understand the spatial pattern and uncertainty of soil C and N and their relationship with vegetation and soil attributes, as well as develop effective approaches to estimate soil C and N over large landscapes and regions. The goal of this study was to improve our understanding of the spatial pattern of soil organic carbon (SOC) and total nitrogen (TN) and their controlling factors in savanna landscapes and develop efficient sampling strategies for evaluating the effects of woody invasion. Specific objectives of this study were to: (1) Quantify the spatial pattern and uncertainty associated with SOC and develop efficient sampling strategies to estimate SOC storage; (2) Assess the influence of soil and vegetation factors on spatial distribution of SOC and TN; and (3) Determine the influence of physical variables related to landscape position and soil on woody vegetation structure. Conditional sequential indicator simulations indicated that woody encroachment into grassland increased both spatial heterogeneity and uncertainty of SOC, which increased errors in estimating SOC storage. Stratified random sampling with higher density in woody patches, plus structured sampling in cluster with strong spatial pattern, substantially increased estimation accuracy. Efficient sampling strategies for estimating SOC storage were developed based on these findings. Direct and spatial correlation and scaling analyses showed that SOC and TN were strongly correlated with litter and root biomass. Invaded woody vegetation has the most impact on spatial distribution of SOC and TN. Canonical correspondence analysis showed that variables related to landscape position were the primary factors determining the spatial distribution of woody species. These new insights will facilitate the estimation of soil C and N pools at landscape and regional scales, and will help evaluate the potential impacts of woody plant encroachment on the biogeochemistry of C and N.