No description available.

No description available.

No description available.

published_or_final_version / Geography / Doctoral / Doctor of Philosophy

<p>Environmental attitudes have interested scholars for decades, but researchers have insufficiently appreciated the low salience of the environment, and the enormous complexity of this issue area. In this dissertation, I investigate how these features influence the way ordinary citizens think about the environment.</p><p>Research into the dynamics of public opinion has found a generic relationship between policy change and public demands for activist government. Yet, less is known about the relationship between policy and attitudes in individual issue areas. In the first chapter, I investigate the influence of a variety of factors on public opinion in a particularly complex policy area---the environment. To study the short-run and long term dynamics of environmental public opinion, I generate an annual metric of environmental attitudes running from 1974 to 2011. Consistent with prior research, I find the economy and major environmental disasters play an important role in aggregate environmental opinion. However, actual policy innovations are found to play only a limited role in attitude formation. Instead, the party label of the president appears to affect demand for environmental activism, when other factors are held constant.</p><p>Scholarly research has found a weak and inconsistent role for self-interest in public opinion, and mixed evidence for a relationship between local pollution risks and support for environmental protection. In the second chapter, I argue that focusing events can induce self-interested responses from people living in communities whose economies are implicated by the event. I leverage a unique 12-wave panel survey administered between 2008 and 2010 to analyze public opinion toward offshore oil drilling before and after the Deepwater Horizon oil spill. I find that residence in counties highly dependent upon the offshore drilling industry was predictive of pro-drilling attitudes following the spill, though not prior to the spill. In addition, there is no significant evidence that residence in a county afflicted by the spill influenced opinion. This chapter concludes that local support for drilling often arises only after focusing events make the issue salient.</p><p>Previous research into place effects has provided mixed evidence about the effect of geography on public opinion. Much of the work finding a relationship is susceptible to methodological criticisms of spuriouness or endogeneity. In the third chapter, I leverage a unique research design to examine the influence of residential setting on environmental attitudes regarding water use. The findings indicate that local drought conditions increase individuals' level of concern about the nation's water supply. In addition, drought conditions are related to public attitudes towards water use regulation, with those living in drought-afflicted counties more likely to support government regulation. This chapter provides a firm foundation for research attempting to demonstrate that local conditions have a causal effect on public opinion.</p> / Dissertation

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Opinion dynamics in a society of interacting agents may lead to consensus or to the coexistence of different opinions. The interplay between social network change and opinion formation is complex, because the agents, their social interactions and the changing social structure over time, are themselves complex. DeGroot proposed a prescriptive model for achieving consensus, where agents revise their opinions at each time step by taking a weighted average of the opinions of neighbours. This thesis contains three main contributions. First, we introduce a generalisation of the DeGroot model and examine the long-time behaviour of the model, with and without insistent agents. Second, we consider opinion formation on networks which are themselves dynamic, where the dynamics may be completely random or based on homophily and triadic closure. The weights that agents place on the opinions of neighbours are also dynamic, based on a rule where weights decrease with increased difference in opinions. Third, we examine the effect of a sudden, temporary or permanent shift in the opinions of some agents. Two dynamics are considered for the network change over time; random switching (RS) network dynamics, and homophily and triadic closure (HT) network dynamics. We prove that the RS network dynamics enhances consensus formation and network connectivity, compared to the HT network dynamics where we show by simulation that different opinions can persist. We investigate the in uence of the presence of a minority of insistent agents and prove that for a connected static network, insistent agents with the same opinion in uence the final opinions to converge to their own opinion, thus leading to consensus. In contrast, lack of consensus persists when insistent agents have different opinions. This conclusion also holds for the RS network dynamics model. However, for the HT network dynamics model, coexistence of different opinions can persist even when insistent agents have the same opinion. This finding regarding the HT dynamics is of particular interest as it relates to observations in the real-world. We also investigate the in uence of a sudden shift in the opinions of some agents on the outcome of final opinions. The case of either a temporary shift in opinions or a permanent shift in opinions is examined. Additionally, the in uence of the time of the introduction of a shift, the number and the network positions of initial recipients of the shift in opinions is investigated. The overall effect of an opinion shift is measured by its in uence on the stabilisation time of the final opinions.

With the growing popularity and availability of opinion-rich resources such as social media platforms and networks, new opportunities arise as people can now share their opinions and also seek or understand the opinion of others about a specific topic or event. This growth has fuelled interest in opinion mining which seeks to understand opinions, attitudes, judgements and evaluations with respect to an entity or its aspects. The proliferation of reviews, ratings and online expressions have turned into a valuable asset to businesses seeking to manage their reputation, market their products, or identify new opportunities through opinion analysis. On the side of consumers, opinion mining serves as an information source that can support decision making. In this research, we focus on some fundamental challenges in opinion mining and make three contributions. First, we develop a curated corpus for training and evaluating opinion mining models. This corpus annotates sentiment and topic information at both sentence and review levels. It also captures the sentiment and topic time-variance information of the reviews. We demonstrate through experiments that this dataset supports opinion mining tasks such as contrastive opinion mining, and joint sentence and document level sentiment and topic analysis. As the corpus has a time-variance characteristic, it could also support studies in sentiment/topic dynamic analysis. Second, we propose a model for mining contrastive opinion from textual data (contraLDA). Unlike existing models that require input data to be separated into different collections beforehand, contraLDA models contrastive opinion from both single and multiple text collections. The model can also be flexibly trained in weakly-supervised and fully-supervised settings. In addition, the contraLDA model not only mines contrastive opinion but also quantifies the strength of opinion contrastiveness towards the topic of interest. The contraLDA model extracts relevant sentences related to the topics, making sentiment-bearing topics more interpretable. Third, we present an aspect extraction method which integrates a Natural Language Processing (NLP) algorithm and word embedding model to identify implicit and explicit aspect expressions from texts. Unlike existing systems, the proposed approach also maps aspect expressions to their corresponding aspect categories. This process allows easy identification of sentences about different aspects of a product. We demonstrate that this unsupervised approach is comparable to state-of-the-art models.

Class of 2007 Abstract / Objectives: The purpose of this study is to determine the reason for the concerns expressed by the nurses in two wards of the University Medical Center. This information will be used to identify areas for improvement in the Pyxis procedures. The objectives of the proposed project include the following: 1) To conduct a descriptive prospective study based on the responses made by nurses about the Pyxis machines, and 2) To determine the most common concerns. Methods: The study was a two phase descriptive prospective study of data obtained through two surveys. The first survey was formative in nature to determine the general area that is the source of the concerns, and the second survey was performed to determine specific opinions that the nurses have about the system. The subjects were nurses that use the Pyxis machines in the 4 Northeast and 4 Northwest wards at the University Medical Center during the time the survey was given. The sample size was approximately 45 subjects. This is the estimated amount of nurses that work in the two wards that were studied. All nurses surveyed used the Pyxis system in one of the two wards. Use of the machines was defined as performing any of the procedures that correspond to the system i.e. dispensing, doing returns, wastes, drawer failure recovery, narcotic inventory, discrepancy documentation, and reporting medication errors. The primary dependent variables were the opinions the nurses have about the Pyxis machine system and what specific factors they have concerns with. Demographic variables included age, gender, how long the subjects have worked in the ward, how many hours per week they work, how often they use the Pyxis machine, degree of education, and what shift they work. A preliminary survey was done to determine the general aspect of the system that produces the most concerns. The survey was distributed at the monthly staff meeting in paper form and then presented to the nurses personally while they were working on the floor. The data from the first survey was then reviewed and used to produce a second survey designed and implemented to discover more detailed and specific opinions of the system. The second survey was also distributed in person to the nurses while they were working on the floor. Results: Analysis of the data from the first survey showed that the concerns expressed by the nurses concerns were primarily based around the stocking procedures of the Pyxis system. The second survey showed that the Pyxis machine should be stocked twice a day at 6:00am and 6:00pm, the finger print scanner has caused nurses some problems, and that ZofranTM should be available for override. Conclusions: The study showed that in order to optimize the capability of the Pyxis system, it should be stocked twice a day at 6:00am and 6:00pm with the enough of the appropriate medications to last until the next restock. According to the comments made by the nurses, there have also been significant problems with the finger print scanner, and not being able to override ZofranTM is also an issue. If possible and plausible, addressing these issues would lead to a smoother work flow, a better working environment, and better outcomes for the patients.

Not available