There is interest in deriving case-mix adjusted standardised mortality ratios so that comparisons between healthcare providers, such as hospitals, can be undertaken in the controversial belief that variability in standardised mortality ratios reflects quality of care. Typically standardised mortality ratios are derived using a fixed effects logistic regression model, without a hospital term in the model. This fails to account for the hierarchical structure of the data - patients nested within hospitals - and so a hierarchical logistic regression model is more appropriate. However, four methods have been advocated for deriving standardised mortality ratios from a hierarchical logistic regression model, but their agreement is not known and neither do we know which is to be preferred. We found significant differences between the four types of standardised mortality ratios because they reflect a range of underlying conceptual issues. The most subtle issue is the distinction between asking how an average patient fares in different hospitals versus how patients at a given hospital fare at an average hospital. Since the answers to these questions are not the same and since the choice between these two approaches is not obvious, the extent to which profiling hospitals on mortality can be undertaken safely and reliably, without resolving these methodological issues, remains questionable.

Few studies have quantified the effects on academic performance; none has investigated, as this study does, the effects of immigration, home language, and school mobility on academic development over time. What makes this study unique is its melding of sociological and psychometric perspectives – an approach that is still quite new. Logistic regression was used to analyze data from Ontario’s 2007-2008 Junior (Grade 6) Assessment of Reading, Writing and Mathematics, with linked assessment results from three years earlier, to investigate students’ academic achievement. The focus of this study is on whether the students maintained proficiency between Grades 3 and 6 or achieved proficiency in Grade 6 if they were not proficient in Grade 3. The results indicate that Grade 3 proficiency is the strongest predictor of Grade 6 proficiency and that home language or interactions with home language are also significant in most cases. In addition, students who speak a language other than or in addition to English at home are, in general, a little more likely to be proficient at Grade 6. Most students who were born outside of Canada were significantly more likely than students born in Canada to stay or become proficient in Reading, Writing, and Mathematics by Grade 6. These results highlight the importance of considering the enormous heterogeneity of immigrants’ experiences when studying the effects of immigration on academic performance.

Few studies have quantified the effects on academic performance; none has investigated, as this study does, the effects of immigration, home language, and school mobility on academic development over time. What makes this study unique is its melding of sociological and psychometric perspectives – an approach that is still quite new. Logistic regression was used to analyze data from Ontario’s 2007-2008 Junior (Grade 6) Assessment of Reading, Writing and Mathematics, with linked assessment results from three years earlier, to investigate students’ academic achievement. The focus of this study is on whether the students maintained proficiency between Grades 3 and 6 or achieved proficiency in Grade 6 if they were not proficient in Grade 3. The results indicate that Grade 3 proficiency is the strongest predictor of Grade 6 proficiency and that home language or interactions with home language are also significant in most cases. In addition, students who speak a language other than or in addition to English at home are, in general, a little more likely to be proficient at Grade 6. Most students who were born outside of Canada were significantly more likely than students born in Canada to stay or become proficient in Reading, Writing, and Mathematics by Grade 6. These results highlight the importance of considering the enormous heterogeneity of immigrants’ experiences when studying the effects of immigration on academic performance.

Material addicional: http://hdl.handle.net/10230/20652 / En aquesta tesi es duu a terme una anàlisi de la implantació de la terminologia en l’ús amb l’objectiu de trobar les causes que expliquen l’èxit o el fracàs de les propostes denominatives. Partim de la premissa que existeix un conjunt de factors que influencien positivament o negativament la implantació terminològica. En aquest treball estudiem la implantació en l’ús de la terminologia normalitzada de l’àmbit de la informàtica i les tecnologies de la informació i la comunicació (TIC) en un corpus especialitzat. Establim una metodologia de treball que ens permet d’obtenir dades quantitatives i qualitatives per a la recerca de factors d’implantació i, a continuació, ens proposem d’estudiar-ne alguns, que tenen a veure amb aspectes lingüístics (brevetat de les denominacions i acostament formal i semàntic amb el terme original), socio¬lingüístics (àmbit d’ús de la terminologia, via d’entrada dels termes en un àrea temàtica determinada i concurrència terminològica) i procedimentals (difusió dels termes a través de diccionaris generals i especialitzats). L’anàlisi ens proporciona dades empíriques per confirmar que alguns dels aspectes àmpliament tractats en la bibliografia són efectivament factors d’implantació. / This work presents a terminology implantation analysis with the aim to find the causes explaining the success or failure of terms in real language use. We depart from the premise that there are factors which influence terminology implantation in a positive o negative way. Our study analyses the implantation of standardised terms of Computer Science and Information and Communication Technologies (ICT) in a specialised corpus. We establish a specific methodology in order to obtain quantitative and qualitative data for the research of implantation factors. In this work we study some of these factors, especially those related with linguistic aspects (brevity of terms and formal and semantic proximity to the original term), sociolinguistic aspects (field of use, entry of terms in a specific domain and terminological competition) and procedural aspects (term dissemination through general and specialised dictionaries). The results obtained provide empirical data to confirm some of the aspects widely discussed in the literature as real implantation factors.

Dialectal Arabic speech research suffers from the lack of labelled resources and standardised orthography. There are three main challenges in dialectal Arabic speech recognition: (i) finding labelled dialectal Arabic speech data, (ii) training robust dialectal speech recognition models from limited labelled data and (iii) evaluating speech recognition for dialects with no orthographic rules. This thesis is concerned with the following three contributions: Arabic Dialect Identification: We are mainly dealing with Arabic speech without prior knowledge of the spoken dialect. Arabic dialects could be sufficiently diverse to the extent that one can argue that they are different languages rather than dialects of the same language. We have two contributions: First, we use crowdsourcing to annotate a multi-dialectal speech corpus collected from Al Jazeera TV channel. We obtained utterance level dialect labels for 57 hours of high-quality consisting of four major varieties of dialectal Arabic (DA), comprised of Egyptian, Levantine, Gulf or Arabic peninsula, North African or Moroccan from almost 1,000 hours. Second, we build an Arabic dialect identification (ADI) system. We explored two main groups of features, namely acoustic features and linguistic features. For the linguistic features, we look at a wide range of features, addressing words, characters and phonemes. With respect to acoustic features, we look at raw features such as mel-frequency cepstral coefficients combined with shifted delta cepstra (MFCC-SDC), bottleneck features and the i-vector as a latent variable. We studied both generative and discriminative classifiers, in addition to deep learning approaches, namely deep neural network (DNN) and convolutional neural network (CNN). In our work, we propose Arabic as a five class dialect challenge comprising of the previously mentioned four dialects as well as modern standard Arabic. Arabic Speech Recognition: We introduce our effort in building Arabic automatic speech recognition (ASR) and we create an open research community to advance it. This section has two main goals: First, creating a framework for Arabic ASR that is publicly available for research. We address our effort in building two multi-genre broadcast (MGB) challenges. MGB-2 focuses on broadcast news using more than 1,200 hours of speech and 130M words of text collected from the broadcast domain. MGB-3, however, focuses on dialectal multi-genre data with limited non-orthographic speech collected from YouTube, with special attention paid to transfer learning. Second, building a robust Arabic ASR system and reporting a competitive word error rate (WER) to use it as a potential benchmark to advance the state of the art in Arabic ASR. Our overall system is a combination of five acoustic models (AM): unidirectional long short term memory (LSTM), bidirectional LSTM (BLSTM), time delay neural network (TDNN), TDNN layers along with LSTM layers (TDNN-LSTM) and finally TDNN layers followed by BLSTM layers (TDNN-BLSTM). The AM is trained using purely sequence trained neural networks lattice-free maximum mutual information (LFMMI). The generated lattices are rescored using a four-gram language model (LM) and a recurrent neural network with maximum entropy (RNNME) LM. Our official WER is 13%, which has the lowest WER reported on this task. Evaluation: The third part of the thesis addresses our effort in evaluating dialectal speech with no orthographic rules. Our methods learn from multiple transcribers and align the speech hypothesis to overcome the non-orthographic aspects. Our multi-reference WER (MR-WER) approach is similar to the BLEU score used in machine translation (MT). We have also automated this process by learning different spelling variants from Twitter data. We mine automatically from a huge collection of tweets in an unsupervised fashion to build more than 11M n-to-m lexical pairs, and we propose a new evaluation metric: dialectal WER (WERd). Finally, we tried to estimate the word error rate (e-WER) with no reference transcription using decoding and language features. We show that our word error rate estimation is robust for many scenarios with and without the decoding features.

The purpose of this bachelor’s degree work was to analyse a packaging station in a global manufacturing company in Sweden, and to provide an answer on why it is important with standardised work in preventing wrong deliveries. The purpose of the analysis was to come up with various improvement proposals in order to reduce and prevent wrong deliveries. The analysis of the packaging station was conducted by observing the station and understanding how the workers work and what kind of work procedure do they follow. The observations made it possible to conduct several conclusions on why the problems occurred, one of these conclusions was that the work procedure is not standardised which in turn increased the errors in the station. Through a clear problem description, it was possible to come up with improvement proposals by answering the following research questions:  Main RQ: Why is a standardised workplace important in a manufacturing company? To be able to answer the main research question, the question has been divided into two smaller sub-questions: 1.     How can a packaging area be standardised by using lean process improvement tools? 2.     How can operating standards improve the efficiency in a packaging station? The approach applied in this study and project has been qualitative research with an inductive approach, by reviewing scientific articles and observing the work in the packaging station it was possible to analyse and compare the collected data in order to come with a conclusion to the problem. The data that was collected in form of interviews has been continuously analysed in order to have as relevant data as possible for the improvements work.  By comparing the collected data from the literature and the case findings it was possible to lay a foundation for the analysis where the proposed improvements came from and to conclude a final proposed improvements that reduces the wrong deliveries in the packaging station. In conclusion, standardised work is important in order to prevent wrong deliveries because it contributes to a more clear and sustainable structure that the workers follow. By having standardised instructions and methods that is being followed, the amount of errors is reduced in the packaging station.  Wrong deliveries are reduced and prevented when there is a standardised work procedure that is based on the lean ideology in a packaging station

Learning to read is a complex and demanding skill which is vital in order for children to be able to access a broad curriculum of learning within the school environment. Reading requires the integration of many different processes, it is possible that difficulties with one or more of these processes has the possibility to interfere with reading ability. The research aimed to investigate the presence and co-occurrence of difficulties across many factors thought to be involved in the reading process. Data were collected from 126 schoolchildren, aged 8-10 years on performance measures associated with reading; reading ability, visual sensory and oculomotor function, visual perception, attention, memory, phonological awareness and rapid naming. Differences in mean performance between different reading ability groups (ANOVA), and correlations between the variables studied, were used to investigate the presence and magnitude of any relationships. Many of the variables studied were found to be significantly different between reading ability groups and significantly correlated with reading ability to varying degrees. The analysis of multiple single-case studies determined that each child has a unique pattern of strengths and weaknesses and that many children including ‘average/above average’ readers, show below average performance on several measures included in the study, with affected skills rarely existing in isolation. Thus, it is recommended that an individualised multi-factorial approach is taken to the assessment of children struggling to read. This will require communication by a multi-professional team to ensure all possible contributing factors are explored to enable each child to achieve their potential. / College of Optometrist / Some material in this thesis is unavailable for copyright reasons.

Risk-adjustment schemes are used to monitor hospital performance, on the assumption that excess mortality not explained by case mix is largely attributable to suboptimal care. We have developed a model to estimate the proportion of the variation in standardised mortality ratios (SMRs) that can be accounted for by variation in preventable mortality. The model was populated with values from the literature to estimate a predictive value of the SMR in this context-specifically the proportion of those hospitals with SMRs among the highest 2.5% that fall among the worst 2.5% for preventable mortality. The extent to which SMRs reflect preventable mortality rates is highly sensitive to the proportion of deaths that are preventable. If 6% of hospital deaths are preventable (as suggested by the literature), the predictive value of the SMR can be no greater than 9%. This value could rise to 30%, if 15% of deaths are preventable. The model offers a 'reality check' for case mix adjustment schemes designed to isolate the preventable component of any outcome rate.

Le comportement de tout être humain sain d’esprit s’explique par une multitude de raisons. Outre la science psychologique, le droit, y compris le droit des contrats, s’intéresse à ces raisons, appelés motifs. Cet intérêt résulte du fait que le comportement des contractants n’est pas abstrait ; il révèle toujours l’existence de divers motifs. Or, le droit objectif ne peut accorder une importance juridique à tous les motifs des parties, étant donné que les motifs représentent une catégorie psychologique et que leur prise en compte illimitée mettrait en péril la sécurité juridique. La présente thèse de doctorat a pour but de proposer une analyse critique de la prise en compte des motifs des parties à partir d’une comparaison des droits français, serbe et anglais / The behaviour of every individual of sound mind is explained by a variety of reasons. Besides psychological science, Law, including contract law, is interested by those reasons, called motives. This interest results from the fact that the behaviour of contracting parties is not abstract; it always reveals the existence of a variety of motives. However, the legal rules of a given domestic system cannot attribute an importance to all motives of parties. This is so because motives represent a psychological category and taking into account in an unlimited way those motives would imperil security in law. This present doctoral thesis aims at proposing a critical analysis of the taking into account of parties’ motives through a comparative study of French, Serbian and English law.

Aluminium die casting is a process used to transform molten aluminium material into automotive gearbox housings, wheels and electronic components, among many other uses. It is used because it is a very efficient method of achieving near net shape with the required mechanical properties. Life Cycle Assessment (LCA) is a technique used to determine the environmental impacts of a product or process. The Life Cycle Inventory (LCI) is the initial phase of an LCA and describes which emissions will occur and which raw materials are used during the life of a product or during a process. This study has improved the LCI technique by adding in manufacturing and other costs to the ISO standardised methods. Although this is not new, the novel application and allocation methods have been developed independently. The improved technique has then been applied to Aluminium High Pressure Die Casting. In applying the improved LCI to this process, the cost in monetary terms and environmental emissions have been determined for a particular component manufactured by this process. A model has been developed in association with an industry partner so this technique can be repeatedly applied and used in the prediction of costs and emissions. This has been tested with two different products. Following this, specialised LCA software modelling of the aluminium high pressure die casting process was conducted. The variations in the process have shown that each particular component will have different costs and emissions and it is not possible to generalise the process by modelling only one component. This study has concentrated on one process within die casting but the techniques developed can be used across any variations in the die casting process.