This paper analyzes the two test forms used by TOEFL: IBT and PBT. The analysiswill focus on the reading comprehension section, its design features, content, andscoring results. The aim is to assess the reliability of the two test forms as well as toidentify factors influencing candidate performance in the reading comprehension test.Three factors are identified: test setting, test difficulty and scoring methods and results.The latter two will be focused on because test difficulty consistency directly decides thetest result consistency. And as the goal of the candidate is to achieve as high a score aspossible, and success is measured in terms of numbers, score reliability is a primaryconcern for both candidate and examining body alike.

Indiana University-Purdue University Indianapolis (IUPUI) / This research project evaluated four reading projects and their authentic materials. The projects were used in a U.S. university ESP program for post-bachelor Afghan students. By utilizing field-related American biographies, online academic articles, various news articles, field-specific article databases, and short fiction for in-depth reading and writing projects, the Afghan students’ reading performances proved and disproved the relevance and importance of these authentic materials as ways to improve their IBT TOEFL reading scores and academic reading proficiencies related to their future work at a U.S. graduate school. Data were collected and evaluated from three TOEFL scores and an in-depth, post program survey.

The holistic listening comprehension of 112 Kyoto University students, operationalized as TOEFL iBT listening (long listening), was investigated with a battery of 12 tests, including a phoneme and word recognition test, a test of short 10-second listening, a test of long 3- to 5-minute listening, a reading comprehension of listening scripts test, listening and reading cloze tests, a gap-filling test designed to assess syntactic awareness, a grammatical error detection test, and the Vocabulary Size Test. Rasch analyses were employed to yield person ability measures; these measures were used for correlation studies, a series of linear regression analyses, principal components analysis, and structural equation modeling. Long listening correlated most strongly with the reading comprehension test (.756) and the listening cloze test (.705), and these two variables explained as much variance in long listening as all the variables combined in a linear regression (68%). Of the two prominent components yielded by a principal components analysis, capturing sounds and processing for meaning, long listening loaded significantly only on processing for meaning (.727) and showed no notable loading on capturing sounds. When long listening comprehension was viewed as a two-stage activity consisting of capturing input and processing that input for meaning, the participants were found to rely mainly on processing for meaning. As a result, long/holistic listening had more in common with reading comprehension than with short listening, for which the first stage of input capture was more important. As a part of this study, long listening was expressed as a product of aural word recognition and processing for meaning as in the Simple View of Reading, where reading comprehension is regarded as a product of decoding and linguistic comprehension. While the Simple View of Reading typically accounts for 48% of the variance in reading comprehension, its listening counterpart in this study explained up to 58% of the variance; as much as an improved version of the Simple View of Reading named the Component Model of Reading. The identification of the structural equation models required an additional component for a total of three latent variables; availability of written text, aural activities, and processing for meaning. The three-latent-variable model for long listening incorporated all the variables as indicators except for the grammatical error detection due to its insignificant contribution to holistic understanding. Generally speaking, structural equation approach produced models which were in good qualitative agreements with correlation studies, principal components analysis, and multiple regression; thus, providing an integrative view and a unified treatment of the participants' proficiency with a focus on long listening. Overall, the results highlighted the importance of processing for meaning, a skill largely shared with reading comprehension, for the long listening comprehension of Kyoto University students. This finding indicates a transfer of meaning formation skill from L1 and L2 reading to L2 listening. / CITE/Language Arts

Made available in DSpace on 2016-04-28T18:22:57Z (GMT). No. of bitstreams: 1 Renata Mendes Simoes.pdf: 1583071 bytes, checksum: b2dfb5b69b92ceadf64ea28788255342 (MD5) Previous issue date: 2014-11-26 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / This research aims at investigating how an English for Specific Purpose one-to-one course, mainly focused on the speaking skill, meets the students learning needs and prepares them to take the TOEFL iBT test (Test of English as a Foreign Language Internet-based Test). The study is grounded on the theoretical principles about English for Specific Purposes as supported by Hutchinson and Waters (1987); Dudley-Evans and St. John (1998); Ramos (2004, 2009); Belcher (2009); Basturkmen (2010); needs analysis according to Hutchinson and Waters (1987); Dudley-Evans and ST. John (1998); Long (2005); task-based teaching as proposed by Willis (1996); Willis and Willis (2001, 2007); Ellis (2003, 2005); Skehan (2003); Nunan (2004); course design and material development according to Hutchinson and Waters (1987); Dudley Evans and St. John (1998); Graves (2000); Tomlinson (2001, 2003); and language assessment as defended by McNamara (2000) and Scaramucci (2004, 2009, 2011). Methodologically, this research is considered a case study (STAKE, 1998; JOHNSON, 1992), and it made use of the following instruments for data collection: questionnaires in the beginning of the course to identify the students profiles, and at the end of the course to learn about the students perceptions as to their progress; audio recording of two speaking tasks on the first and last day of class; teacher-researcher s diaries at the end of every class containing the students perceptions of their performance in class. The results revealed the students satisfaction regarding the course methodology and material, as well as the perception of reading and listening skills improvement. It was interesting to note that, although the main focus of the course was on the speaking skill, the perception of improvement of the writing skill was higher than that of the speaking skill. The students narratives also indicated the importance of teacher-student interaction and praised the attention given by the teacher to their emotional aspects. This work aims at bringing contributions not only to ESP course teachers and to those who work with language assessment, but also to the research field as it lacks studies related to the speaking skill and private classes / Este trabalho tem por objetivo investigar como um curso de Inglês para Fins Específicos, com priorização na produção oral, realizado por meio de aulas particulares, atende as necessidades de aprendizagem de alunos e os capacita a prestar o exame TOEFL iBT (Test of English as a Foreign Language Internet-based Test). O referencial teórico englobou os conceitos sobre Inglês para Fins Específicos de acordo com Hutchinson e Waters (1987); Dudley-Evans e St. John (1998); Ramos (2004, 2009); Belcher (2009); Basturkmen (2010); análise de necessidades conforme proposto por Hutchinson e Waters (1987); Dudley-Evans e ST. John (1998); Long (2005); ensino baseado em tarefa segundo Willis (1996); Willis e Willis (2001, 2007); Ellis (2003, 2005); Skehan (2003); Nunan (2004); desenho de curso e desenvolvimento de materiais à luz de Hutchinson e Waters (1987); Dudley Evans e St. John (1998); Graves (2000); Tomlinson (2001, 2003); e avaliação de línguas conforme defendido por McNamara (2000) e Scaramucci (2004, 2009, 2011). Este trabalho, caracterizado como estudo de caso (STAKE, 1998; JOHNSON, 1992), utilizou para coleta de dados os seguintes instrumentos: questionários no início do curso para investigar o perfil do aluno, e ao final do curso, para identificar as percepções dos alunos quanto a seu progresso; gravação em áudio no primeiro e no último dia de aula de duas tarefas de produção oral; diários da professora-pesquisadora ao final de cada aula com as percepções do aluno referente ao desempenho em aula. Os resultados revelaram a satisfação dos alunos quanto à metodologia e material didático empregados no curso e a percepção da melhora na compreensão oral e escrita. Foi interessante notar que, embora o curso tivesse mantido o foco principal na produção oral, a percepção de melhora da produção escrita foi superior à melhora da produção oral. Os relatos dos alunos também indicaram a importância da interação professora-aluno e a atenção dada pela professora ao aspecto emocional dos alunos para essa percepção de melhora. Este trabalho pretende trazer contribuições não só para aqueles que lecionam em cursos de Inglês para Fins Específicos e trabalham com avaliações de língua, mas também para o campo de estudos por ser este ainda carente de pesquisas sobre aulas particulares e produção oral

Dans ce mémoire, nous examinons certaines propriétés des représentations distribuées de mots et nous proposons une technique pour élargir le vocabulaire des systèmes de traduction automatique neurale. En premier lieu, nous considérons un problème de résolution d'analogies bien connu et examinons l'effet de poids adaptés à la position, le choix de la fonction de combinaison et l'impact de l'apprentissage supervisé. Nous enchaînons en montrant que des représentations distribuées simples basées sur la traduction peuvent atteindre ou dépasser l'état de l'art sur le test de détection de synonymes TOEFL et sur le récent étalon-or SimLex-999. Finalament, motivé par d'impressionnants résultats obtenus avec des représentations distribuées issues de systèmes de traduction neurale à petit vocabulaire (30 000 mots), nous présentons une approche compatible à l'utilisation de cartes graphiques pour augmenter la taille du vocabulaire par plus d'un ordre de magnitude. Bien qu'originalement développée seulement pour obtenir les représentations distribuées, nous montrons que cette technique fonctionne plutôt bien sur des tâches de traduction, en particulier de l'anglais vers le français (WMT'14). / In this thesis, we examine some properties of word embeddings and propose a technique to handle large vocabularies in neural machine translation. We first look at a well-known analogy task and examine the effect of position-dependent weights, the choice of combination function and the impact of supervised learning. We then show that simple embeddings learnt with translational contexts can match or surpass the state of the art on the TOEFL synonym detection task and on the recently introduced SimLex-999 word similarity gold standard. Finally, motivated by impressive results obtained by small-vocabulary (30,000 words) neural machine translation embeddings on some word similarity tasks, we present a GPU-friendly approach to increase the vocabulary size by more than an order of magnitude. Despite originally being developed for obtaining the embeddings only, we show that this technique actually works quite well on actual translation tasks, especially for English to French (WMT'14).

No description available.

The purpose of this study was to determine which types of student application information, as well as demographic information obtained through a questionnaire after matriculation, best predicted later academic performance in an international English-medium university in Japan, and to examine the "big picture" of how cognitive and non-cognitive variables interact over time in accounting for student success in an English for Academic Purposes (EAP) program and in the regular university program. The study was divided into three parts that separately examined student application information, university entrance examinations, and the larger picture of student success. In the first part of the study, a hierarchical multiple regression was employed to determine the extent to which a variety of variables derived from application information predicted grade point average (GPA) in the EAP program, as well as first-year GPA and final GPA in the regular university program. The independent variables examined in the main regression analysis were: high school grade point average (HSGPA); ITP TOEFL scores obtained in April of the students' first year at the university; and hensachi rankings of the students' high schools. Results indicated that HSGPA was a consistently significant predictor of all levels of university GPA. ITP TOEFL scores significantly predicted EAP GPA, and <italic>hensachi / CITE/Language Arts

No description available.

No description available.

No description available.