This thesis defends objections derived from current high stakes educational summative assessment arrangements. The objections are derived from considering consequences of the philosophical problem of vagueness. I argue that current arrangements unsuccessfully address the issue of vagueness but have done so fully recognising that vagueness is no trivial threat to assessment practice. The issue stands alongside other issues considered of equal significance to fairness and equality as other key issues addressed in educational assessment literature such as reliability and validity. Vagueness raises peculiar problems about borderlines that are typically understood through consideration of the standard puzzle of vagueness, the sorites. The puzzle asks why, if ten thousand grains of sand make a heap, one grain doesn't? Although deceptively simple and of ancient pedigree, this puzzle has continued to prove difficult to solve and there is no settled account amongst contemporary philosophers as how best to conceptualise the phenomenon. Pressed into service of educational assessment, we can adapt the sorites to ask: 'If 100 marks is a pass, why isn't one mark?' The problem is that wherever one decides to draw a sharp boundary between pass and fail at any particular mark, there seems no justification for not drawing it at just a slightly lower or slightly higher one. Vagueness threatens the system because it provides reasons for thinking that no one should fail. If the puzzle is set up in reverse, it also provides reasons for thinking that no one should pass. I presuppose that any assessment system unable to offer principled rational discriminations of value is incompetent. The sorites carries this threat. My thesis- presses a concern that, because of Cresswell's powerful advocacy of a certain approach to vagueness, the educational assessment community has considered his perspective to be the last word. Unfortunately this is not the case. I draw attention to the fact that there are several options and that different consequences for assessment follow depending on which approach is adopted. I argue that the epistemic approach to the issue solves the problem of vagueness using a classical logic framework used in intrasubjective judgment in science. I argue that Roy Sorensen's epistemic solution to the sorites satisfies best the requirements of an assessment systems' need for similar intrasubjective judgment.

No description available.

No description available.

No description available.

No description available.

No description available.

Following a survey of the history and nature of the cloze test as applied to second-language assessment, the study looks at the role of introspective and verbal-report data in understanding second-language test-taking processes. Particular attention is paid to the verbal report task format known as ‘think-aloud’. The theoretical bases of this procedure are critiqued, and some problematic aspects of its use in the study of linguistic tasks are discussed. Attention is drawn to apparent divergences between the model of think-aloud and its real-world applications. The use of think-aloud in the study of cloze test-taking by German and Japanese first-language informants is discussed, and a number of specific shortcomings identified. These shortcomings lie mainly in the areas of practical sample-size, interpretability and comprehensiveness of data, and negative affective responses on the part of informants. A modification to the ‘classical’ think-aloud is proposed, labelled non-continuous reporting, and the results of this method are compared to those of think-aloud. It is concluded that the advantages of non-continuous reporting outweigh its shortcomings. A further alternative real-time data-gathering procedure is proposed, the so-called ‘annotated cloze’, and its strength and drawbacks discussed. The relative efficiencies of annotated cloze and the two variants of think aloud are examined in terms of their ability to generate a picture of how test-takers process cloze passages, and suggestions are offered regarding optimal use of these task formats in the elicitation of further data.

No description available.

This study aims to provide a comprehensive view of the national educational system, and then to determine the values of some internationally-used indicators within Oman's schools, and their effects upon student performance in the 12 th grade examinations. From this analysis, the relationships between each variable and the optimal values for overall improvements in the total examination marks of students in the 12 th grade are derived. In pursuing this aim, the research conducts the first statistical analysis of ten educational indicators using data gathered by the Omani Ministry of Education. A logistic regression method is adopted to study repetition, drop-out, and promotion rates, using data pertaining to the school years 2006/2007-2007/2008. The analysis in this respect includes all governmental schools' students in grades 1-12. A linear mixed model is then used to explore data for the school years 2005/2006-2007/2008, relating to class size, school size, student/teacher ratio, teacher/class ratio, student age, student gender, and school gender mix. For these indicators, the analysis includes all the 12th grade students. The results reveal that girls are more likely to promote to the following grades than are boys, since they have low drop-out and repetition rates. This phenomenon causes the average student age of 12th grade girls to be less than that of boys. Grade 12 emerges as having the highest dropout rate, and grade 5 has the highest repetition rate. The distribution of the total marks of the 12 th grade students displays a long left tail, reflecting the fact that many students have low total marks. School size and class size are identified as having a negative effect upon the total examination mark. As the school size increases, so too do classes, but they are not allowed to become too large. The student age has the greatest significant effect on student performance among all the variables, student gender coming second in this ranking. Students who are young for their academic year are performing better than those who are older for the year, and girls perform better than boys. The research concludes that the basic education system in the Sultanate of Oman is in good health but that room for development and improvement does exist.

This submission draws on research from twelve publications, all addressing some aspect of the broad research question: “Can interactive computer-marked assessment improve the effectiveness of assessment for learning?” The work starts from a consideration of the conditions under which assessment of any sort is predicted to best support learning, and reviews the broader literature of assessment and feedback before considering the potential of computer-based assessment, focusing on relatively sophisticated constructed-response questions, and on the impact of instantaneous, tailored and increasing feedback. A range of qualitative and quantitative research methodologies are used to investigate factors which influence the engagement of distance learners of science with computer-marked assessment and computer-generated feedback. It is concluded that the strongest influence on engagement is the student’s understanding of what they are required to do, including their understanding of the wording of assessment tasks and feedback. Clarity of wording is thus important, as is an iterative design process that allows for improvements to be made. Factors such as cut-off dates can have considerable impact, pointing to the importance of good overall assessment design, and more generally to the power and responsibility that lie in the hands of remote developers of online assessment and teaching. Four of the publications describe research into the marking accuracy and effectiveness of questions to which students give their answer as a short phrase or sentence. Relatively simple pattern-matching software has been shown to give marking accuracy at least as good as that of human markers and more sophisticated computer-marked systems, provided questions are developed on the basis of responses from students at a similar level. However, educators continue to use selected-response questions in preference to constructed-response questions, despite concerns over the validity and authenticity of selected-response questions. Factors contributing to the low take-up of more sophisticated computer-marked tasks are discussed. E-assessment also has the potential to improve the learning experience indirectly by providing information to educators about student engagement and student errors, at either the cohort or individual student level. The effectiveness of these “assessment analytics” is also considered, concluding that they have the potential to provide deep general insight and an early warning of at-risk students.