Social buying is a recent marketing innovation in which provides Pareto-improving welfare gains to merchants, consumers, and brokers. Consumers benefit from access to significant discounts on advertised products and services, the broker benefits from taking a significant cut in each transaction with very low fixed costs, and merchants are able to reduce their advertising costs, gain access to new markets and drive traffic to their stores. The phenomenal growth of social buying carries commensurate risks for brokers, including increased competition due to a lack of service differentiation and low entry barriers. The complete social buying transaction is completed over two stages: the initial online e-commerce transaction and the subsequent fulfilment transaction where the voucher is redeemed with the merchant. In order to explore the sustainability of the social buying business model, it is necessary to identify the factors which drive loyalty behaviours in social buying, as well as the interrelationships between the factors. This research proposes from the marketing literature Oliver’s (1980) expectancy-disconfirmation theory (EDT) as the main theoretical framework on which to model these relationships. EDT is then successfully synthesised with DeLone and McLean’s (2003) information systems success model to create a framework which can appropriately model both the online and traditional stages of the social buying transaction. This study contributes to the marketing literature by establishing EDT as a suitable framework for investigating social buying. It is believed that this study is the first to do so. Furthermore, it is believed this is the first study examining the social buying innovation in the South African context. / Graduate School of Business Leadership / MBA

Eccles' expectancy-value model of achievement motivation suggests that beliefs about ability and expectations for success are a strong predictor of grades and differences in task value underlie differences in motivation and achievement. This model has not been previously investigated in the context of high stakes examinations in the UK and this study therefore explores the relationships between expectations, values and A-level achievement in 930 students. This is important given the significance of these qualifications for future life pathways. Furthermore, studies examining the subjective task value (STV) patterns across school subjects, rather than domain specific ones, are rare, highlighting the additional importance of this work. A mixed methods design was used. A questionnaire collected information on a student's background (SES, gender, ethnicity), the expectations and STV attached to A-levels, and their future and general life expectations and values. Some of these relationships were also explored using 20 semi-structured interviews. The qualitative data illustrated that studying A-levels confirmed aspects of students' identity but also facilitated changes to their goals and academic skills, having positive effects, contrary to the argument that high stakes assessment has a negative impact on individuals. Unsurprisingly parents and teachers were perceived to be influential. As predicted, expectations and values were related to A-level achievement. As there is a lack of research into the effects of these variables on A-level outcomes these findings are valuable. Eccles' original three factor model of STV could not, however, be supported. In this research the utility construct was removed. Further exploration of the STV construct is warranted. Socio-economic status was positively related to both achievement and expectations about achievement. Girls had lower expectations but placed higher value on their A-levels. There were, however, no gender differences in achievement. Employing the expectancy-value model in this UK context has been useful in explaining the motivational patterns underlying A-level qualifications and the findings have implications for enhancing outcomes and narrowing educational gaps in this student population.

Despite the vast research on the productive aspect of rewards, little is known on how the changes in employees’ behaviour, made to enhance their chances of achieving a reward, influence employee well-being. Previous work has failed to address the process of reward-seeking behaviour from an employee’s point of view as the focus was on the motivational aspect of rewards. This thesis uses the case of Malta’s financial institutions to examine the relationship between reward-seeking behaviour from bonuses and promotions and employee well-being by drawing on expectancy theory and self-determination theory. To achieve its aims, this study adopts a qualitative approach, wherein 42 semi-structured interviews with employees and four interviews with human resources managers are conducted at financial institutions in Malta – two of which are small and medium-sized enterprises and one is a large-sized institution. Memos and diary notes are also used to complement the data collected from the semi-structured interviews. Overall, the results strongly support the idea that while almost everyone values rewards, employees differ in their willingness to engage in reward-seeking behaviour and its influence on well-being. This thesis contributes to knowledge through the development of a theoretical model – the four quadrant reward-seeking behaviour – well-being model. This typology based model classifies employees into four main categories, namely, highly motivated, apathetic, work-life balanced and work-life imbalanced. This two by two matrix also led to another model that depicts reward-seeking behaviour and well-being as a non-sequential process. The findings have practical implications for human resources practitioners as they now have the capacity to visualise the actual employee mix according to the categories of the model and act on any significant gaps.

Modelos preditivos estimam o risco de eventos ou agravos relacionados à saúde e podem ser utilizados como ferramenta auxiliar em tomadas de decisão por gestores e profissionais de saúde. Algoritmos de machine learning (ML), por sua vez, apresentam potencial para identificar relações complexas e não-lineares presentes nos dados, com consequências positivas na performance preditiva desses modelos. A presente pesquisa objetivou aplicar técnicas supervisionadas de ML e comparar sua performance em problemas de classificação e de regressão para predizer respostas de interesse para a saúde pública e a medicina. Os resultados e discussão estão organizados em três artigos científicos. O primeiro apresenta um tutorial para o uso de ML em pesquisas de saúde, utilizando como exemplo a predição do risco de óbito em até 5 anos (frequência do desfecho 15%; n=395) para idosos do estudo \"Saúde, Bem-estar e Envelhecimento\" (n=2.677), segundo variáveis relacionadas ao seu perfil demográfico, socioeconômico e de saúde. Na etapa de aprendizado, cinco algoritmos foram aplicados: regressão logística com e sem penalização, redes neurais, gradient boosted trees e random forest, cujos hiperparâmetros foram otimizados por validação cruzada (VC) 10-fold. Todos os modelos apresentaram área abaixo da curva (AUC) ROC (Receiver Operating Characteristic) maior que 0,70. Para aqueles com maior AUC ROC (redes neurais e regressão logística com e sem penalização) medidas de qualidade da probabilidade predita foram avaliadas e evidenciaram baixa calibração. O segundo artigo objetivou predizer o risco de tempo de vida ajustado pela qualidade de vida de até 30 dias (frequência do desfecho 44,7%; n=347) em pacientes com câncer admitidos em Unidade de Terapia Intensiva (UTI) (n=777), mediante características obtidas na admissão do paciente à UTI. Seis algoritmos (regressão logística com e sem penalização, redes neurais, árvore simples, gradient boosted trees e random forest) foram utilizados em conjunto com VC aninhada para estimar hiperparâmetros e avaliar performance preditiva. Todos os algoritmos, exceto a árvore simples, apresentaram discriminação (AUC ROC > 0,80) e calibração satisfatórias. Para o terceiro artigo, características socioeconômicas e demográficas foram utilizadas para predizer a expectativa de vida ao nascer de municípios brasileiros com mais de 10.000 habitantes (n=3.052). Para o ajuste do modelo preditivo, empregou-se VC aninhada e o algoritmo Super Learner (SL), e para a avaliação de performance, o erro quadrático médio (EQM). O SL apresentou desempenho satisfatório (EQM=0,17) e seu vetor de valores preditos foi utilizado para a identificação de overachievers (municípios com expectativa de vida superior à predita) e underachievers (município com expectativa de vida inferior à predita), para os quais características de saúde foram comparadas, revelando melhor desempenho em indicadores de atenção primária para os overachievers e em indicadores de atenção secundária para os underachievers. Técnicas para a construção e avaliação de modelos preditivos estão em constante evolução e há poucas justificativas teóricas para se preferir um algoritmo em lugar de outro. Na presente tese, não foram observadas diferenças substanciais no desempenho preditivo dos algoritmos aplicados aos problemas de classificação e de regressão analisados. Espera-se que a maior disponibilidade de dados estimule a utilização de algoritmos de ML mais flexíveis em pesquisas de saúde futuras. / Predictive models estimate the risk of health-related events or injuries and can be used as an auxiliary tool in decision-making by public health officials and health care professionals. Machine learning (ML) algorithms have the potential to identify complex and non-linear relationships, with positive implications in the predictive performance of these models. The present research aimed to apply various ML supervised techniques and compare their performance in classification and regression problems to predict outcomes of interest to public health and medicine. Results and discussion are organized into three articles. The first, presents a tutorial for the use of ML in health research, using as an example the prediction of death up to 5 years (outcome frequency=15%; n=395) in elderly participants of the study \"Saúde, Bemestar e Envelhecimento\" (n=2,677), using variables related to demographic, socioeconomic and health characteristics. In the learning step, five algorithms were applied: logistic regression with and without regularization, neural networks, gradient boosted trees and random forest, whose hyperparameters were optimized by 10-fold cross-validation (CV). The area under receiver operating characteristic (AUROC) curve was greater than 0.70 for all models. For those with higher AUROC (neural networks and logistic regression with and without regularization), the quality of the predicted probability was evaluated and it showed low calibration. The second article aimed to predict the risk of quality-adjusted life up to 30 days (outcome frequency=44.7%; n=347) in oncologic patients admitted to the Intensive Care Unit (ICU) (n=777), using patients\' characteristics obtained at ICU admission. Six algorithms (logistic regression with and without regularization, neural networks, basic decision trees, gradient boosted trees and random forest) were used with nested CV to estimate hyperparameters values and to evaluate predictive performance. All algorithms, with exception of basic decision trees, presented acceptable discrimination (AUROC > 0.80) and calibration. For the third article, socioeconomic and demographic characteristics were used to predict the life expectancy at birth of Brazilian municipalities with more than 10,000 inhabitants (n=3,052). Nested CV and the Super Learner (SL) algorithm were used to adjust the predictive model, and for evaluating performance, the mean squared error (MSE). The SL showed good performance (MSE=0.17) and its vector of predicted values was used for the identification of underachievers and overachievers (i.e. municipalities showing worse and better outcome than predicted, respectively). Health characteristics were analyzed revealing that overachievers performed better on primary health care indicators, while underachievers fared better on secondary health care indicators. Techniques for constructing and evaluating predictive models are constantly evolving and there is scarce theoretical justification for preferring one algorithm over another. In this thesis no substantial differences were observed in the predictive performance of the algorithms applied to the classification and regression problems analyzed herein. It is expected that increase in data availability will encourage the use of more flexible ML algorithms in future health research.

A model proposed by Holroyd and Coles (2002) stating that error related negativity (ERN) is caused by a decrease in mesencephalic dopamine output to the ACC was tested. A group of individuals with Parkinson's disease (N = 16) and an age and education matched group free from neurological disorder (N = 16) completed a card guessing task where the magnitude of monetary penalties and rewards for incorrect and correct answers was varied by block. Individuals with Parkinson's disease were tested after an overnight washout from dopamingeric medications. The amplitude of the mid-frontal negativity elicited by feedback was analyzed with spatial and temporal principal components analyses. Dipole source analyses were also performed. Analyses revealed no significant differences in the mid-frontal negativity amplitude between the two groups. In addition, the magnitude of consequence and the validity of response had no significant effects on fERN amplitude although there was a trend for higher magnitude consequences to be associated with larger fERN amplitude. Dipole analyses indicated the source of the mid-frontal negativity fell into the cingulate, specifically the cingulate gyrus. The results suggest that the mid-frontal negativity elicited by feedback indicating an error was made remains intact in individuals with Parkinson's disease. This does not support predictions made by Holroyd and Coles' model in regard to this group unless disruptions to the system that produces the fERN do not occur until later stages in the disease. An additional finding was a late positive potential for the error trials which began approximately 450 milliseconds after feedback and continued throughout the epoch. The ramifications of this wave are discussed.

Using the Health and Retirement Study Panel core files from 1996 to 2014, I analyze how Medicare Part D impacted access to prescription drug coverage by various demographic factors such as race, gender, and income. In Chapter 1, I find the highest take-up rates for those who were white, female, and with higher incomes. However, increases in coverage were high across the board, such that Medicare Part D also improved drug insurance coverage for those who were black, male, and with lower income. Thus, although Medicare Part D did increase prescription drug insurance coverage for seniors across the board, I also find potential for improvement in enrollment for difficult-to-reach groups. Next, Chapter 2 examines the impact of Medicare Part D on mortality. Although I do not find an impact on the life expectancy of respondents as a whole, I do find a significant positive effect for black respondents, indicating that Medicare Part D may have mattered more for disadvantaged groups. The largest impact is for black men, who have an additional 9 percentage point chance of living to age 73 for an additional 8 years of coverage (significant at the 5% level). When looking only at cardiovascular mortality, which is more likely to be influenced by drug coverage, I find improvements in life expectancy for the total population, with stronger effects for minorities and men. Overall, my findings suggest that Medicare Part D did move the needle on its goal: to improve the health of those who, without government intervention, had the most difficulty paying for prescription drugs. Chapter 3 looks at the impact of Medicare Part D prescription drug coverage on cost-related medication adherence, food insecurity, and finances among seniors. It would be reasonable to assume that Medicare Part D, which led to near-universal drug coverage among senior citizens, could allow seniors to shift money previously spent on drug expenditures to other areas. The strongest effect of Medicare Part D is on cost-related medication nonadherence, leading to a 21% decrease for an additional 8 years of Medicare Part D coverage. The impact is even stronger for the black male population (30%). I fail to reject the null hypothesis that Medicare Part D did not reduce food insecurity or household debt. Overall, Medicare Part D appears to have improved the financial stability of seniors.

The understanding of why students decide to become music therapists is valuable information for music therapy educators and policy makers and published information is lacking. The use of expectancy-value theory provides a framework to understand student choices. Researchers can better understand why students purse a career in music therapy by examining students’ abilities, beliefs, expectancies, and values related to the choice. The purpose of this study was to examine why current undergraduate and equivalency students want to be music therapists. Music therapy students (N = 129) throughout the United States provided insight into their decision to become music therapists by completing a survey with questions about their expectancies and values. Results indicated that both undergraduate and equivalency students tended to choose to be music therapists because they expected music therapy to be a career that requires hard work and expert knowledge while maintaining their interest and morale. Results of a correlational analysis suggest there are several choice-making variables that may influence each other, indicating that the decision to become a music therapist is a complex process for students. Finally, the results of a Mann-Whitney U test suggested that there was no significant difference between the expectancies and values that influenced undergraduate and equivalency students’ choices to become music therapists. Music therapists in many capacities may use this information to improve recruitment and engagement of music therapy students.

Numerous studies from various research groups have already shown the usefulness of alcohol expectancies as predictors of long-term future alcohol consumption. The present study extends this line of research by directly testing whether alcohol expectancies measured in the moment using free association are useful as predictors of alcohol consumption in the next few hours. An ecological momentary assessment (EMA) procedure was used to examine how alcohol expectancies might fluctuate during days in which many people expect to drink (e.g. Fridays, Saturdays) and how these fluctuations in alcohol expectancies might predict future drinking and/or co-vary with important contextual variables during that same day. The results supported our main hypothesis that increases in positively-valenced alcohol expectancies would be observed a few hours to minutes before engaging in alcohol consumption. These findings provide further evidence that anticipatory information processing is a key part of the motivational system that directs future behavior, and that probing expectancies in real-time can be useful for predicting alcohol consumption in the near future.

<p>There is today much evidence for a positive relationship between alcohol and violence. There are however still many questions about the nature of the relationship. Somewhat simplified, the research on the link between alcohol and violence can be divided into four different lines of research, research focusing on: the effect of alcohol as a psychoactive substance, the drinking context, the personality of the drinker, and societal attitudes, expectations and values. The dissertation focuses on the last area, the importance of the cultural context. One influential theory within this field is the time-out theory formulated by MacAndrew and Edgerton in the late 1960s. Since drunken comportment varied between cultures and between different contexts within the same culture and changed over time, they suggested that the effects of alcohol on people’s behaviour was socially constructed. They suggest that many societies had a created a time-out situation for drunken behaviour that explained people’s behaviour while drunk.</p><p>The general aim for this dissertation is to study young Swedes’ attitudes, experiences and expectations around drunken behaviour, with a special focus on expectancies around alcohol as a cause and excuse for violence. Three different data sets have been used. The first study is based on 4 focus-group interviews with Swedish football fans during the European football championship in Holland in 2000. The second material is eight focus-group interviews involving 47 students aged 18-20 living in Stockholm. The last material is a nationally representative survey of young adults, 16-25 years old. One part of the survey consisted of 4 vignettes which we used to elicit cultural norms around drunken behaviour.</p><p>The findings suggest that young Swedes believe that alcohol can be used as a means to accomplish a pleasurable state of mind, and that alcohol could be used as a means to transgression – since alcohol reduce inhibitions it could be used to put them in a less controlled mode. The rationale for those changes was often described in terms of the psychoactive effect of alcohol. However, it was also shown that the context was important. When the situation demanded alcohol the most, their expectations together with the situation almost turned water into beer. It was also shown that there was a norm which said that one should not use alcohol as an excuse, but on the other hand, the participants said that they used alcohol as an excuse and that they thought that it was accepted. Alcohol could work as an excuse since alcohol made the aggressor look less deviant and the acts less severe. The vignette studies indicated that an aggressor who was drunk when he committed a violent act was seen as less blameworthy than a sober or less drunk aggressor. However, this applied only under certain circumstances: alcohol seemed to be a better excuse if the victim is drunk as well and the act is relatively severe.</p><p>Taken together, the studies suggest that the Swedish drinking culture provides people with a drunken excuse, which helps young people to expand the room for possible action.</p>

<p>Bandura definierar (1997) outcome expectancy som individens skattning av sannolikheten att ett beteende ska leda till ett specifikt utfall. Utöver individens skattning av sitt eget beteendes följder (individual outcome expectancy), är hennes bedömning av sin grupps möjlighet att nå ett specifikt utfall (collective outcome expectancy) en del av begreppet (Riggs & Knight, 1994). Studien syftar till att undersöka huruvida arbetsrelaterad- self efficacy, locus of control och collective efficacy predicerar outcome expectancy. En enkätundersökning genomfördes med 102 deltagare från olika yrkesgrupper. Resultatet visade att arbetsrelaterad- self efficacy och locus of control samvarierade med individual outcome expectancy och att collective efficacy predicerade collective outcome expectancy. Dock var gruppstorleken av betydelse för prediktionen. Studiens hypoteser bekräftades av tidigare forskning.</p>