Pesquisas com bilíngues e usuários de línguas adicionais utilizam diferentes medidas para avaliar proficiência no processo de seleção dos seus participantes. A proficiência linguística de um indivíduo pode ser medida através de testes de proficiência, testes de desempenho e medidas de autoavaliação. Questionários de histórico da linguagem contêm perguntas sobre a experiência que os indivíduos têm com as suas línguas e escalas de autoavaliação de proficiência. Estudos sugerem que usuários de línguas adicionais são capazes de reportar a sua proficiência de forma consistente com o seu desempenho em medidas padronizadas (MARIAN et al., 2007; LUK et al., 2013). Nesse contexto, a presente dissertação teve como objetivo analisar a proficiência autoavaliada de usuários de inglês como língua adicional através do Questionário de Experiência e Proficiência Linguística (QuExPLi). Para isso, foram desenvolvidos dois estudos. No primeiro estudo, correlacionamos as variáveis da experiência linguística – marcos de idade, tempo de imersão e uso atual da língua – de 535 participantes com a proficiência reportada por eles no Questionário de Experiência e Proficiência Linguística (QuExPLi), a fim de investigar quais fatores se associariam à proficiência autoavaliada. No segundo estudo, correlacionamos as mesmas variáveis da experiência com o inglês como língua adicional com os resultados no TOEFL ITP de 112 participantes, que fizeram parte do primeiro estudo, para investigar quais fatores se associariam à proficiência demonstrada em um teste de proficiência padronizado. Além disso, correlacionamos a proficiência autoavaliada no questionário com os resultados no TOEFL ITP, para averiguar a relação entre duas diferentes medidas de proficiência e reunir evidências de validade do QuExPLi. Os resultados sugerem que, em relação às variáveis da experiência linguística, os fatores relacionados a marcos de idade, tempo de imersão e uso atual da língua associam-se com a proficiência auto reportada pelos participantes. Porém, dentre esses fatores, apenas os meses passados em um país em que o inglês é falado correlacionaram-se de forma significativa com a proficiência dos participantes demonstrada através do TOEFL ITP. A discrepância entre os fatores associados à proficiência autoavaliada e aqueles que se associaram com o resultado no TOEFL ITP pode advir da natureza acadêmica do teste. Os resultados também revelam uma correlação positiva significativa moderada entre a proficiência autoavaliada no QuExPLi e a pontuação dos participantes no TOEFL ITP, sugerindo que os participantes, usuários de inglês como língua adicional, conseguem reportar sua proficiência de forma acurada. / Research with bilinguals and additional languages learners use different measures to evaluate proficiency while selecting participants for studies. An individual‟s language proficiency can be measured through proficiency tests, performance tests or self-evaluation measures. Language background questionnaires include questions about individuals‟ experience with their languages and self-assessment proficiency measures. Studies suggest that users of additional languages can report their proficiency in a way that is consistent with their performance in standard measures (MARIAN et al., 2007; LUK et al., 2013). In this context, the goal of the present thesis was to analyze self-assessed proficiency of users of English as an additional language through the Questionário de Experiência e Proficiência Linguística (QuExPLi). In order to do that, two studies were designed. In the first study, we correlated language experience variables – age, time of immersion and current use of the language - of 535 participants with their self-reported proficiency on the QuExPLi, in order to investigate which factors would associate with self-assessed proficiency. In the second study, we correlated the same language experience variables with the scores on the TOEFL ITP of 112 individuals, who also participated in the first study, in order to investigate which factors would associate with the proficiency demonstrated in a standardized proficiency test. Moreover, we correlated participants‟ self-assessed proficiency with their scores on the TOEFL ITP, in order to investigate the relationship between two distinct measures of proficiency and gather evidences of validity of the QuExPLi. Results suggest that the factors related to age, time of immersion and current use of the language are associated with participants‟ self-assessed proficiency. However, only the months spent in a country where English is spoken correlated significantly with participants‟ scores on the TOEFL ITP. The discrepancy between the factors associated with self-reported proficiency and those associated with the results on the TOEFL ITP can be derived from the academic nature of the test. Results also show a moderate positive correlation between participants‟ self-assessed proficiency and their scores on the TOEFL ITP was found, suggesting that users of English as an additional language are able to report their proficiency accurately.

Internationally English is increasingly being used as a language of instruction in education. This is also true within the South African context, a country with eleven official languages. Many non-native English speaking (NNES) students, for whom English might be their second, third or even fourth language, are studying through the medium of English. Previous studies on this topic acknowledge that the level of English proficiency which NNES students have, affects their academic performance. The problem under investigation in this case study was the relationship between students’ academic performance in particular modules and how their written responses in examinations contributed to them failing a module. The purpose was to establish to which extent limited English proficiency contributed to the poor academic performance of NNES preservice teachers studying through distance education. This study is underpinned by Cummins’ theory of Basic Interpersonal Communication Skills and Cognitive Academic Language Proficiency (1984). It is supported by related literature which emphasises the challenges NNES students experience in understanding academic content and responding to written assessment tasks in a language other than their home language. A quantitative approach was used which focused on the written responses to examination papers of thirty undergraduate B Ed students who had failed a particular module, at a private higher education institution. The contribution of inadequate or incoherent English, as the reason why marks were not allocated to answers, was calculated and analysed. Results indicated that students’ English grammatical proficiency does have an influence on their academic performance, though it is not the main contributing factor to students failing their modules. For this case study it was determined that the pre-service teachers’ inadequate English written proficiency contributed almost a third (25,6%) of the reasons for their poor academic performance. This finding suggests the need to develop new teaching strategies to accommodate and offer language support to NNES students in higher education institutions that offer qualifications using English as the medium of instruction. / Dissertation (MEd)--University of Pretoria, 2017. / Humanities Education / MEd / Unrestricted

This study explores the effect of an explicit training module on improving language learner ability to accurately self-assess second-language proficiency. There was a total of 409 intermediate and advanced level participants across six languages: French, Italian, Japanese, Portuguese, Russian and Spanish. Most of the participants had extensive immersion backgrounds. The Language Ability Self Evaluation Resource (LASER) was used to measure the perceived proficiency of the participants (self-assessed language ability). The Oral Proficiency Interview – Computer (OPIc) was used to measure the actual proficiency of the participants (rater-assessed language ability). The participants were randomly assigned to either the control or experimental group. Both groups started and ended the self-assessment the same way, with the experimental group receiving the training directly before the self-assessment questions. The training module was designed to include clarification, instruction, modeling, practice and immediate feedback while only adding 10 minutes to the overall assessment time. The limited time was dictated by the overall desire for maximum usability within already busy curricula and by already busy students. All participants completed the OPIc within 30 days of taking the LASER. The results were unexpected, revealing that those learners who received the training were more likely to over-assess their own proficiency level than learners who received no training at all. Possible reasons for this outcome are explored and possible ways to improve the assessment tool are discussed.

L2 proficiency gains during study abroad vary widely across individuals and programs, and much of the research in the study abroad literature attempts to identify the causes of this variance. Social network data has proven useful in explaining some of the variance in oral proficiency gains (Baker-Smemoe, Dewey, Bown, & Martinsen, 2014; Isabelli-García, 2006), and the current study builds on those findings by applying the same methodology to listening and reading proficiency in addition to speaking. Proficiency gains in listening, reading, and speaking were measured for 17 students from a US university studying abroad in Nanjing, China for one semester. Social network measures focused on interaction with native speakers (NS) were taken at the beginning, middle, and end of the study abroad program using the Study Abroad Social Interaction Questionnaire. Linear regression analyses showed that social network measures accounted for nearly 46% of the variance in listening gains, nearly 82% of the variance in reading gains, and nearly 46% of the variance in oral proficiency gains. These findings make a strong case for applying social network methods to understand listening and reading proficiency gains in study abroad.

Elicited imitation (EI) is a method of assessing oral proficiency in which the examinee listens to a prompt and attempts to repeat it back exactly as it was heard. Research over recent decades has successfully established correlation between EI testing and other oral proficiency tests, such as the Oral Proficiency Interview (OPI) and the OPI by computer (OPIc). This paper details the history of oral proficiency assessment as well as that of EI. It then outlines the development process and validation of a Portuguese Elicited Imitation test. The processes of item selection and item validation are detailed followed by the criterion-related validation through a statistical correlation analysis of participants' results on an official American Council on the Teaching of Foreign Languages (ACTFL) OPIc and their predicted OPIc scores which were based on their results of the Portuguese EI calibration test. Results of the statistical analysis revealed a strong correlation between the predicted scores of the EI test and the actual OPIc scores. In order to go beyond previously completed EI research, this paper addresses the issue of face validity which has been a challenge for the proliferation of EI testing. Analysis of a survey administered after participants' completion of the two tests (OPIc and EI) addresses the experiences and reactions of the participants to the two testing formats. Suggestions for future use of EI as well as future research will be presented.

This study investigated the role of vocabulary size in assessing learners from various L1 backgrounds and for institutional placement. Participants included native speakers of Spanish, Korean, Japanese, and Mandarin, who took a test that estimated the size of their productive English vocabulary. First, the vocabulary size of students from each language group was compared and the relationship between vocabulary size and institutional placement level was examined. Then, scores were analyzed to determine how cognates and loan words influenced the participants' performance on the vocabulary test. Further, students' vocabulary size scores were correlated with their placement scores that evaluated reading, writing, speaking, listening, and grammar. Finally, the vocabulary size of new students was compared to returning students within the same institutional placement level. Within the same institutional placement levels, speakers of Spanish and Korean had larger English vocabularies than speakers of Japanese and Chinese. It was also found that there is a clear distinction between students at varying institutional placement levels in terms of their vocabulary size. The cognate analysis revealed that students learn cognates faster than non-cognates. The correlation analysis revealed that vocabulary size correlated most with speaking, and listening, followed by grammar, then reading and writing. Finally, the cross-sectional analysis indicated that the vocabulary size of newly placed students was generally larger than that of continuing students.

No description available.

No description available.

Student Number: 9109265Y A research report submitted to the Faculty of Humanities, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Master of Arts (Psychology) / The Academic Proficiency Assessment battery evaluates language and study skills. This study focused on the internal consistency reliability and predictive validity of this battery for Information Technology students. In terms of reliability, the Time management, Note-taking skills, and Debilitating stress scale on the achievement anxiety questionnaire were found to be internally consistent. However, the items in the English proficiency, Reading comprehension, Memorisation skills, concentration skills and motivating stress scale on the achievement anxiety questionnaire require modification or replacement. Intercorrelations across questionnaires necessitate further streamlining. In terms of predictive validity, a significant negative relationship was found between Note-taking Skills and academic performance (R2 adj = 8,3%). Matric results remain the best predictor accounting for 11% of the variance in CGPA. Cumulatively, Matric results and Note-taking skills accounted for 13,34% of the variance. None of the biographical variables significantly predicted CGPA. Despite the apparent lack of relationship between individual predictors and the criterion, a discriminant function analysis indicated that all the academic proficiencies, with the exception of English proficiency, correctly predicted pass or failure in 72% of the cases. The lack of relationship between English proficiency and pass/failure can be accounted for in terms of the type of courses studied as part of an Information Technology programme.Overall, the results would suggest that Matric results remain the best predictor of academic performance in Information Technology courses, but at a very low level. Given the lack of reliability in the majority of the subtests of the Academic Proficiency Battery, the use of the APA battery for selection for remedial intervention for Information Technology students is not yet justified.

Published Article / The issue of language of learning and instruction in communities with diverse linguistic and cultural backgrounds, globally and in South Africa, is always shaped by socio-economical, political, ideological and hegemonic factors. The language of instruction and learning in South Africa is politically charged, not only because of colonial legacy, but more profoundly because of the apartheid legacy. This article reports on the results of the pedagogies of teachers teaching in English. The researcher found that most non-native English teachers use their native language pronunciation when teaching English, which results in mother tongue interference. On-native speakers end up transferring articulation habits of first language to second language. There is thus a need to improve quality in pedagogical practices in this diverse South African background since the language of instruction is different from the learners' first language for the majority of the population. The research in this article aims to investigate current pedagogic practices engaged by teachers from different cultural backgrounds when teaching English as a medium of instruction and how their practices affect learners in acquiring knowledge of the English language. The study uses an interpretive paradigm and in particular, adopts social constructivism to embed discussions. The study mainly implements a qualitative approach although the quantitative approach is used to quantify biographical data. The researcher found that two educational systems seem to exist in South African schools: some educators use English as a medium of instruction only, whereas others allow code switching. Education thus becomes a reproductive mechanism of social class differences.