To United Minds, large-scale surveys are an important offering to clients, not least the public opinion poll Väljarbarometern. A risk associated with surveys is satisficing – sub-optimal response behaviour impairing the possibility of correctly describing the sampled population through its results. The purpose of this study is to – through the use of multivariate outlier detection methods - identify those observations assumed to be non-representative of the population. The possibility of categorizing responses generated through satisficing as outliers is investigated. With regards to the character of the Väljarbarometern dataset, three existing algorithms are adapted to detect these outliers. Also, a number of randomly generated observations are added to the data, by all algorithms correctly labelled as outliers. The resulting anomaly scores generated by each algorithm are compared, concluding the Otey algorithm as the most effective for the purpose, above all since it takes into account correlation between variables. A plausible cut-off value for outliers and separation between non-representative and representative outliers are discussed. The resulting recommendation is to handle observations labelled as outliers through respondent follow-up or if not possible, through downweighting, inversely proportional to the anomaly scores.

What is the developmental trajectory of the skills that underlie emergent literacy during the preschool years? Are there individual characteristics which predict whether a child will be at-risk for difficulties in acquiring literacy skills? Does a child’s experience in a high-quality early care and education environment enhance the development of his or her emergent literacy? The present study is an investigation of the individual and environmental factors relevant to children’s emergent literacy skills as they unfold in time. Using a combination of principal components analysis, growth modeling with a multi-level approach, and propensity score analysis, the trajectories of growth in emergent literacy were examined. In addition to child characteristics, the effects of early child environments on emergent literacy were also examined. The effects of home literacy environment and of high-quality early care and education environments were investigated using propensity score matching techniques. The growth in emergent literacy was examined using a nationally representative dataset, the Early Childhood Longitudinal Study – Birth cohort (ECLS-B). Child characteristics, such as primary home language and poverty, were associated with lower initial abilities and suppressed growth in emergent literacy. A high-quality home literacy environment had a strong effect on the growth of children’s emergent abilities, even after controlling for child characteristics. High-quality early care and education environments, as defined by structural attributes of the program such as class size, had a modest impact on the growth of emergent literacy skills for some but not all children. When high-quality early education was defined in terms of teacher interaction, children who are exposed to such care experienced an increase in growth of their emergent literacy abilities. This study provides an examination of individual and group paths toward literacy as an element of school readiness, including the role of environment in the development of literacy skills. These findings have implications for early education policy, especially relevant to state-funded preschool programs and Early Head Start, to provide insight into contexts in which policy and the investment of resources can contribute most effectively to early literacy development.

When researchers are unable to randomly assign students to treatment conditions, selection bias is introduced into the estimates of treatment effects. Random assignment to treatment conditions, which has historically been the scientific benchmark for causal inference, is often impossible or unethical to implement in educational systems. For example, researchers cannot deny services to those who stand to gain from participation in an academic program. Additionally, students select into a particular treatment group through processes that are impossible to control, such as those that result in a child dropping-out of high school or attending a resource-starved school. Propensity score methods provide valuable tools for removing the selection bias from quasi-experimental research designs and observational studies through modeling the treatment assignment mechanism. The utility of propensity scores has been validated for the purposes of removing selection bias when the observations are assumed to be independent; however, the ability of propensity scores to remove selection bias in a multilevel context, in which group membership plays a role in the treatment assignment, is relatively unknown. A central purpose of the current study was to begin filling in the gaps in knowledge regarding the performance of propensity scores for removing selection bias, as defined by covariate balance, in multilevel settings using a Monte Carlo simulation study. The performance of propensity scores were also examined using a large-scale national dataset. Results from this study provide support for the conclusion that multilevel characteristics of a sample have a bearing upon the performance of propensity scores to balance covariates between treatment and control groups. Findings suggest that propensity score estimation models should take into account the cluster-level effects when working with multilevel data; however, the numbers of treatment and control group individuals within each cluster must be sufficiently large to allow estimation of those effects. Propensity scores that take into account the cluster-level effects can have the added benefit of balancing covariates within each cluster as well as across the sample as a whole.

Elder abuse has traditionally been a silent social issue in America. However, with an estimated increase in the older population over the next 50 years, and given the preventable nature of violence, it is quickly becoming a major public health priority area. Each year hundreds of thousands of elderly are abused, neglected, or exploited financially worldwide. In the United States alone, it is estimated that 500,000 cases of elder abuse occur annually—with research indicating that substantiated cases are a mere underreport of the true problem. The US federal government has appointed State Units on Aging to address elder abuse. Georgia’s Division of Aging Services (DAS) is located within the Department of Human Services and administers various services to elderly including advocating for their safety and well being. DAS carries out its work through locally appointed Area Agencies on Aging (AAA). While AAAs serves as a first point of entry for elderly population locally, their involvement in reporting and intervening in elder abuse cases has been limited. The purpose of this capstone project is to examine the AAAs’ stage of readiness to address elder abuse using the Community Readiness Model, developed by researchers at the University of Colorado. Telephone administered surveys were completed with 7 out of the 12 Georgia AAAs. Through a double rater review process, transcripts were coded according to diverse constructs of the Community Readiness Model and ultimately a readiness score was produced. The Community Readiness Score provides insight into evidence-based strategies that can be implemented in order to advance elder abuse intervention and prevention within the AAA communities. The findings from this study provide insights into cost-efficient, tailored strategies that can maximize the use of DAS funding for AAA elder abuse case response and service delivery.

Recommender systems are widely used to achieve a constantly growing variety of services. Alongside with social networks, recommender systems that take into account friendship or trust between users have emerged. In this thesis, we propose an evolution of trust-based recommender systems adapted to decentralized architectures that can be deployed on top of existing social networks. Users profiles are stored locally and are exchanged with a limited, user-defined, list of trusted users. Our approach takes into account friends' similarity and propagates recommendation to direct friends in the social network in order to prevent ratings from being globally known. Moreover, the computational complexity is reduced since calculations are performed on a limited dataset, restricted to the user's neighborhood. On top of this propagation, our approach investigates several aspects. Our system computes and returns to the final user a confidence on the recommendation. It allows the user to tune his/her choice from the recommended products. Confidence takes into account friends' recommendations variance, their number, similarity and freshness of the recommendations. We also propose several heuristics that take into account peer-to-peer constraints, especially regarding network flooding. We show that those heuristics decrease network resources consumption without sacrificing accuracy and coverage. We propose default scoring strategies that are compatible with our constraints. We have implemented and compared our approach with existing ones, using multiple datasets, such as Epinions and Flixster. We show that local information with default scoring strategies are sufficient to cover more users than classical collaborative filtering and trust-based recommender systems. Regarding accuracy, our approach performs better than others, especially for cold start users, even if using less information.

Une désescalade thérapeutique des traitements du cancer du sein est observée notamment grâce à la chimiothérapie néo-adjuvante (CNA). Alors que la technique du ganglion sentinelle (GS) n'est actuellement pas recommandée après CNA, nous avons démontré sa faisabilité dans une série prospective de 129 patientes traitées par CNA avec un taux d'identification de 93,8%, un taux de FN de 14,3% et de 0% pour les patientes N0 avant CNA. En cas d'atteinte duGS, le risque de GNS métastatique était corrélé à la taille tumorale (p=0.016) et la taille de la métastase du GS (p=0.0055). Le nomogramme du MD Anderson (AUC=0.716) et le score de Thenon (AUC=0.778) pouvaient évaluer la probabilité d'atteinte du GNS. L'analyse per opératoire du GS par apposition a permis l'identification d'une métastase chez 72% des patientes (sensibilité=38.2% ; spécificité=97.8%). Les patientes qui présentaient une micro métastase ou la présence de cellules tumorales isolées dans le GS avaient un risque multiplié par 2,3 de FN de l'apposition par rapport aux patientes qui avaient une atteinte macro métastatique. En conclusion, notre travail a permis de montrer que la CNA n'avait pas d'influence négative sur la faisabilité de la technique du GS, en particulier pour les patientes N0 clinique avant traitement. Après CNA, l'analyse per opératoire du GS est possible ainsi que l'utilisation de certains nomogrammes déjà existants pour calculer la probabilité d'atteinte du GNS lorsque le GS est métastatique.

Att vara ideellt engagerad ses av många som en faktor som leder till positiva effekter för individer. Problemet med detta resonemang är att de flesta som engagerar sig aktivt även ärde som redan har dessa fördelaktigaegenskaper.Frågan är därför om det är en kausal effekt eller en stark korrelation som lett till bilden av att ideellt engagemang är positivt för individers livsstandard.Syftet med denna uppsats är att undersöka om ett aktivt engagemang inom ideella organisationer i sig gereffekter hos individerna. Studiengörs med utgång från hypotesen att socialt kapital skapas inom grupper av aktivt ideellt engagerade medlemmar och att detta i sin tur leder till positiva yttringarhos individen.Undersökningen sker genomen kvantitativ longitudinell studie som använder sig av propensity score matchingför att uppskatta den kausala effekten aktivt engagemang kan ha.Studiens resultat indikerar att ingen statistisk signifikant skillnad existerar mellan en person som varitaktiv i en ideell organisationoch en person som ej varit det, vilket tyder på att det aktiva engagemanget inom en ideell organisation i sig inte ger någon effekt. Dessa resultat förkastar därmed hypotesen att aktivt engagemang inom ideella organisationer i sig leder till positiva effekter genom det sociala kapital som skaffas inom denna form av nätverk.

No description available.

Bakgrund: Många företag går i konkurs varje år vilket leder till att olika intressenter drabbas hårt till följd av konkursen. Därmed har konkursprognostiseringsmodeller utvecklats för att ge en tidig varning till intressenterna om företags framtida finansiella kris.  I Sverige använder kreditinstituten sig av sina egna modeller för att förutspå konkurser, dessa modeller är inte publicerade för allmänheten. I och med detta är vi intresserade att tillämpa utländska modeller på den svenska marknaden. Syfte: Syftet med denna uppsats är att undersöka fyra internationella prognostiseringsmodeller för att analysera hur tillämpbara de är på den svenska marknaden. Metod: Undersökningen har baserats på en kvantitativ forskningsstrategi och en deduktiv forskningsansats. Urvalet grundades på de företag som inlett konkurs år 2012 samt en kontrollgrupp bestående av friska företag. Slutligen bestod det slumpmässiga urvalet av 31 konkursföretag och 31 friska företag som tillhör tillverknings- och industribranschen. Teori: Under teoriavsnittet beskrivs de modeller som används i denna studieforskning. Vidare redogörs för nyckeltalens betydelse vid bedömning av företags finansiella förhållanden. Slutligen beskrivs tidigare forskning inom konkursprognostisering. Resultat och slutsats: Modellerna är inte tillämpbara på den svenska tillverknings- och industribranschen då dessa inte har presenterat tillförlitliga resultat på vår studie. Vi anser att en vidare revidering av dessa modeller behövs för att dessa ska kunna tillämpas på den svenska marknaden.

Syftet med denna undersökning är att utveckla den investeringsstrategi som Piotroski (2000) tog fram, grundad på fundamentalanalys, genom att sammanlänka variabler från Lev & Thiagarajan (1993) som visat sig vara värderelevanta indikationer på företags rapporterade resultat. För att genomföra detta utvecklas en modell med Piotroskis (2000) F-score som grund. Antalet signaler i modellen utökas från 9 till 12 stycken. Undersökningen genomförs med data från Stockholmsbörsen under perioden 1998 – 2012. Resultatet visar att både den utvecklade modellen och Piotroskis modell presterar en positiv marknadsjusterad avkastning under hela undersökningsperioden. Samtidigt ökar antalet investeringar i den utvecklade modellen vilket bidrar till en minskad risk och en ökad spridning.