147 Using Tasks to Assess Spanish Language Learning1 Leonardo Herrera Mosquera2* Universidad Surcolombiana, Colombia Abstract The methodology of Task-based teaching (TBT) has been positively regarded by many researchers and language teachers around the world. Yet, this language teaching methodology has been mainly implemented in English as a second language (ESL) classrooms and in English for specific purpose (ESP) courses; and more specifically with advanced-level learners. The present experimental research study aimed at proving the feasibility of a Task-Based Assessment (TBA) approach in a different learning context: A beginning Spanish class. That is to say, contrary to the traditional TBT-TBA implementation, the experiment was conducted in a foreign language class with students bearing a low level of language proficiency. The research produced positive results in terms of students’ linguistic and communicative performance. Keywords: Task, Task-Based Teaching (TBT), Task-Based Assessment (TBA). Resumen La metodología de enseñanza basada en tareas (TBT, por sus siglas en inglés) ha tenido gran acogida por parte de investigadores y profesores alrededor del mundo. Sin embargo, este enfoque ha sido implementado principalmente en programas de inglés como segunda lengua (ESL) o en cursos de inglés con propósitos específicos (ESP); y más enfáticamente con estudiantes de niveles avanzados. El presente trabajo investigativo experimental busca demostrar la viabilidad de una metodología de evaluación basada en tareas (TBA) en un contexto diferente: una clase de español para principiantes. Es decir, contrario a la tradicional implementación de TBT y TBA, este experimento se llevó a cabo en una clase de lengua extranjera con estudiantes de un nivel bajo de competencia lingüística. La investigación arrojó resultados positivos en términos de desempeño comunicativo y lingüístico. Palabras claves: Tareas, Enseñanza basada en tareas (TBT), Evaluación basada en tareas (TBA). 1 Received: August 11th, 2011 / Accepted: May 14th, 2012 2 Email: lehemos@hotmail.com Gist Education and LEarninG rEsEarch JournaL. issn 1692-5777. no. 6, novEmbEr 2012. pp. 147-158 No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 148 Resumo A metodologia de ensino baseada em tarefas (TBT, pelas suas siglas em inglês) tem tido grande aceitação por parte de pesquisadores e professores ao redor do mundo. Entretanto, este enfoque tem sido implementado principalmente em programas de inglês como segunda língua (ESL) ou em cursos de inglês com propósitos específicos (ESP); e mais enfaticamente com estudantes dos níveis avançados. O presente trabalho investigativo experimental busca demonstrar a viabilidade de uma metodologia de avaliação baseada em tarefas (TBA) em um contexto diferente: uma aula de espanhol para principiantes. Melhor dito, contrário à tradicional implantação de TBT e TBA, este experimento foi realizado em uma aula de língua estrangeira com estudantes de um nível baixo de competência linguística. A pesquisa deu resultados positivos em termos de desempenho comunicativo e linguístico. Palavras chaves: Tarefas, Ensino baseado em tarefas (TBT), Avaliação baseada em tarefas (TBA). Literature Review In recent decades, tasks have become an important methodological tool within the language teaching and learning process, especially in the field of Second Language Acquisition (SLA). But what do researchers and teachers mean by task? A frequently cited author on this issue is Long (1985), who offers a general definition that ranges from a non-verbal event such as painting a fence or dressing a child to a more communicative event such as taking a hotel reservation or borrowing a library book. According to Long, almost any action people perform in their daily life can be called a task. This definition has been repeatedly used as a starting reference point by other theorists. Nunan (1989) depicts a more communication-oriented definition stating that a communicative task is “a piece of classroom work which involves learners in comprehending, manipulating, producing or interacting in their target language while their attention is primarily focused on meaning rather than form. The task should also have a sense of completeness, being able to stand alone as a communicative act in its own right”. Breen (1989) calls a task any language learning activity, no matter the level of complexity, that follows a structured workplan. This workplan is intended to provide learners with opportunities to develop usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 149 their linguistic and communicative competences. Breen adds that tasks need to have well-defined objectives, content, procedures, and outcomes in order to facilitate the language learning process. Following on Breen’s perspective of task, we could cite Lee (2000), who offers a definition in which a structured workplan for the task is a requirement. Lee emphasizes two fundamental elements for the accomplishment of a task: there must be a focus on meaning, and interaction must be the means through which the objective is met. Other authors like Bygate, Skehan, and Swain (2001) define tasks as activities that require learners to use language, with emphasis on meaning, to attain an objective. Ellis (2003) defines tasks as activities that call for primarily meaning- focused language use. As we can see from these definitions, tasks call for a focus on meaning so that they resemble linguistic events that occur in real-life circumstances. These real-world tasks are also known as rehearsal or target tasks (Common European Framework, p. 157). Making classroom tasks similar to real-life communicative acts was my main concern when designing assessment tasks for the study herein described. Even if the object of study was a grammatical component, I adapted it in such a way that the assessment task would resemble a real- world communicative event. Teaching methodologies that incorporate such tasks as central components of the curriculum are denominated task-based instruction (TBI) or task-based teaching (TBT). However, the present study focuses on just one portion of the TBT process: Assessment or task-based assessment (TBA). That is, tasks have been designed and implemented in this study as a means to assess students’ linguistic and communicative skills and to offer appropriate feedback. For the purpose of the present study, I designed communicative assessment tasks that combine elements of both types of assessment (system-based and performance-based). The themes, lexical items, and grammatical structures already stated in the Spanish 1 syllabus were incorporated into the design of the tasks in a way that better resembled real-world communicative events, also known as target-language use (TLU) tasks. I also want to add that most of the tasks designed for this study were integrative in that they integrate two or more language skills (listening and writing, reading and speaking, and the like) or two or more linguistic components (grammar, vocabulary, pronunciation, etc.). Methodology Research Design The present experimental research sought to measure the impact of a task-based assessment approach in the learning of Spanish at a middle school level. The target population consisted of all Spanish usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 150 1 students in a middle school in The United States of America. The name of the school will not be disclosed for confidentiality reasons. The accessible population consisted of four classes or groups of Spanish 1 students. These four classes were part of my teaching assignment during the 2009-2010 school year. Since the groups (accessible population) were already established, the experimental group design implemented is the static-group pretest-posttest. An entrance test was administered to both the treatment and control groups, followed by a six-month treatment. An exit test was administered to both groups as well. Differences in performance scores between the treatment and the control groups demonstrate the effect of the treatment considering all the intervening variables and factors. Groups and Sample A sample of convenience was used in this study. Two out of the four Spanish 1 classes were randomly selected as the treatment group and the remaining two classes were designated as the control groups. All four groups consisted of seventh and eighth grade students with ages ranging between 11 and 14 years. The characteristics analyzed in this study were group size, gender, grade level, and Spanish background for both the treatment and the control groups. Students who had had any kind of contact with the Spanish language, whether at home or at school, qualified as having some background in Spanish. This information was obtained from the entrance personal survey. Instrumentation On the first day of class (September 2009), students filled out a form with information regarding personal data, their background in Spanish, and their motivation to learn Spanish. These data provided significant information for the interpretation of results. On the second day, a Spanish entrance test was administered to all four groups (accessible population). This test consisted of 30 questions of vocabulary and grammar. The questions for this test were taken from the textbook online self-tests for the first two chapters. The treatment consisted of 10 task-based tests administered during the first semester, September 2009 to January 2010. These task-based tests were administered only to the two treatment groups. The two control groups were administered traditional quizzes such as multiple-choice, matching, and fill-in-the-blank type tests. The content usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 151 for the task-based tests was dictated by the textbook syllabus and adapted according to real language demands. For example, instead of just asking students to complete drills with verbs in the present tense, students answered a survey of trends and preferences, which required the use of verb inflections in this tense. The process of adapting the tasks according to the Spanish syllabus was completely done by the teacher-researcher. In some task-based models such as that proposed by Breen (1987), students are responsible for performing their own needs analysis, designing procedures, and reflecting on the whole learning process. Candlin (1987), in turn, proposes that teachers may bring curriculum specifications as options for students to consider in their task-designing process. Students in the treatment groups received a complete description of the task (workplan) and started preparing with classmates in the allotted preparation time. The task handouts contained the task description, performance guidelines, and assessment rubric, if necessary. This phase when students learn about the task requirements is what I call pre task. The post task corresponds to the analysis of the task performance by the teacher and the students. By the end of the semester (January 2010), all groups- the treatment and control- took the midterm written exam, which had already been designed by the Foreign Language Department of the school. This test consisted of 139 multiple-choice questions, with 40 listening comprehension questions, 40 vocabulary questions, 40 grammar questions, and 19 textual comprehension questions. Along with the midterm written exam there was a midterm speaking exam. On this speaking exam, all students from both the treatment and the control groups were interviewed by an external interviewer, another Spanish teacher invited to participate in the research. This ten-question interview constituted the tenth assessment task and was titled Entrevista de trabajo 2 (Job Interview 2). Students were questioned about personal information, personality traits, favorite indoor and outdoor activities, school information, and other general topics such as dates and times. All instruments in this study were administered to all groups except for the 10 assessment tasks which corresponded to the treatment. Similarly, all instruments were administered by the classroom teacher (researcher) except for the semester speaking test. The personal information form and the midterm multiple-choice exam were designed by the Foreign Language Department of the school. The ten assessment tasks and the semester speaking test were designed by the researcher. Yet, the speaking test was administered by a guest teacher who had the usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 152 possibility to choose among a wide range of questions—that is, there was not a fixed 10-question interview. It was pivotal to invite an external interviewer for the speaking exam as well as to take the above-mentioned steps in order to make the study more valid. An external teacher or interviewer would not be able to exert any influence on the interviews nor on the results. They would not even know how groups were composed or which groups corresponded to the treatment or control. In this regard Norris et al. (1998) state that validity is one of the main threats to the effective implementation of a Task-Based Language Approach. Data Collection The grading software “Gradequick” was used to enter grades and calculate the groups’ averages (mean and median) for the entrance test and the midterm multiple-choice exam. The semester speaking exam was graded by the tester (guest teacher) using the rubric on the task handout. All ten assessment tasks were graded by the classroom teacher, but these grades were not used for the final analysis of this study. The grading by the teacher-researcher might threaten the validity of the study. The teacher’s journal of observations on the pre task, during, and post task phases were added to the analysis of the quantitative information. Thus, qualitative and quantitative information were combined in the analysis. Data Analysis Scores obtained on the semester written and speaking exams were compared to the scores obtained on the entrance test for each group. Both the group means and the medians (arithmetic measures) were calculated in order to have a more reliable evidence of the results. The comparison between the entrance test and the exit test averages provided relevant information regarding the effect of the treatment (implementation of a task-based assessment approach). Results An entrance test, a midterm speaking exam, and a midterm multiple-choice exam were the principal assessment tools used in the present study to measure the impact of a task-based assessment approach on the learning of Spanish as a foreign language in a middle school in The United States. usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 153 Entrance Test This test was administered on the first week of the 2009-2010 school year (September). It was administered to four Spanish 1 classes in order to measure their initial level of Spanish language proficiency. The test consisted of 30 multiple-choice questions assessing vocabulary and grammar. All four groups obtained scores in the 30-40 range on the entrance test, which meant that all groups were bearing a similar level of language proficiency. These numbers are intended to give us an idea of students’ level of language proficiency and background knowledge before undertaking the treatment for the present study. Midterm Multiple-choice Exam The midterm multiple-choice exam consisted of 139 questions distributed as follows: 40 listening comprehension questions, 35 vocabulary questions, 40 grammar questions, and 24 questions for textual comprehension. This exam was designed by the Foreign Language Department of the school in previous years. Students took this exam on the last week of January 2010. It took them between 50 and 70 minutes to complete this test. The two treatment groups showed the highest averages (83 and 87 respectively), and the spread of the data for all groups was quite similar. Even though the treatment groups were never assessed throughout the semester using multiple-choice type tests, they did very well on this midterm test to the point of obtaining higher scores than the groups who were always assessed using multiple-choice tests. Midterm Speaking Exam The midterm speaking exam consisted of a 10-question interview covering topics such as personal information, likes and preferences, personality traits, school subjects, and other general questions about dates and times. These were the topics covered during the first semester by the four groups. The tester was a Spanish teacher from the school who was invited to assess students in order to make the research more valid. This teacher has great experience teaching Spanish 1 students as well as teaching the Spanish curriculum of the target school. The tester was instructed to select the questions she wanted from a long list, or even to ask the questions she considered suitable for the conversation. These interviews were administered during two 90-minute class periods. usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 154 Once again, the treatment groups obtained the highest averages, 92 and 96 respectively. These results are consistent with the way students in these two groups were assessed throughout the semester. Most of the tasks were aimed at speaking; therefore students were better trained to take this type of test. The difference in averages between the treatment and the comparison groups is both statistically and practically significant, which enables us to suggest the assessment treatment as a great alternative to use in the teaching of Spanish. Yet this suggestion as well as others will be more deeply explained in the following section, along with conclusions. Statistical Analysis This statistical analysis consisted of an independent samples t-test using the statistics program SPSS (Statistical Package for the Social Sciences). What this t-test did was to compare the difference in means obtained in both midterm exams (multiple-choice and speaking) by the treatment and the control groups. From this operation, it was determined if the difference in means was statistically significant. Also calculated was the probability of obtaining the same result by chance. In order to make this conclusion, the .05 level of significance was used. That is, if the difference in means was of .05 or less, it would be considered statistically significant. If the difference in means was bigger than .05, the probability of obtaining the same result by chance is bigger and therefore the research hypothesis may not be well-supported. The statistical result showed that the difference in means was not statistically significant at the 0.05 level. That is, the significance (2 tailed) obtained (.391) is bigger than the 0.05 reference level. In other words, the possibility of obtaining the same difference in means by chance is 39.1 out of 100 (too large probability). However, as previously stated, this difference in means has a practical significance taking into account that the treatment groups were never assessed through multiple-choice type tests during the semester. Yet, they obtained a two-point higher score in their means. On the speaking test, the mean for the treatment groups is 4.5 points higher than the mean for the comparison groups. The statistical operation showed that the difference in means was statistically significant at the 0.05 level. That is, the significance (2 tailed) obtained (0.024) is smaller than the 0.05 reference level. In other words, the possibility of obtaining the same difference in means by chance is 2.4 usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 155 out of 100 (a small probability). This difference was expected because the treatment groups were mostly assessed through oral type tests during the entire semester. To sum up, the treatment groups obtained higher scores in both semester exams, but only in the speaking exam was the group score (means) considered statistically significant according to the independent samples t-tests. Conclusions and Recommendations The treatment applied in this study produced positive results in students’ performance on both semester exams. Even though students in the treatment groups were never assessed through multiple-choice type tests, they obtained higher scores on the midterm multiple-choice exam than the control groups. This result indicates that students can be taught a language, in this case Spanish, following a communicative approach and assessed through communicative tasks, and still be prepared for psychometric type tests (i.e. multiple-choice tests). The main conclusion that this study depicts is that task-based language assessment in a middle-school Spanish class is not only possible but also effective. Students were taught and assessed following a task-based language approach and when they were assessed at the end of the semester, their scores in both the speaking and the written multiple-choice exams were higher than those of the control groups. The language assessment approach implemented in this study is highly recommended in terms of high achievement on speaking tests and on multiple-choice type tests. In other words, the six-month treatment applied in the two Spanish classes at this middle school produced higher levels of communicative and linguistic performance in students. However, for further studies in this specific area (TBA in a foreign language class), it is suggested to have a more extensive period of teaching and assessment in order to confirm the results of this treatment. If future research studies confirm these results, the implementation of a TBA can be generalized to other levels of language proficiency as well as to other language learning processes, as foreign or second language. If TBA is to be adopted, I propose the following recommendations that can make task-based language assessment smoother and the learning process more effective. It was observed that students get very nervous when taking oral tests or performing oral tasks, especially those in which they are to speak in front of the class. Then, it is highly recommended that teachers do significant preparation on the pre task stage so that students’ level of anxiety can be lowered and consequently their language performance can increase in spontaneity and fluency. usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 156 That is why on the pre task stage students need to be given sufficient time to comprehend the task, to ask for clarifications, to practice, and to rehearse interactions. According to Krashen (1985), once the anxiety variable is controlled and the motivation is enhanced, students are better prepared for language learning success (Affective Filter Hypothesis). The post-task is another learning momentum for both teachers and students. On the one hand, the teacher can benefit from students’ reflection and feedback (metacognitive reflection) and make adjustments to the task or the task process. On the other hand, students can benefit from the teacher’s feedback regarding linguistic, communicative, or interactional performance. These two task stages will undoubtedly help students lessen their level of stress as well as reflect on their own language learning process. Group tasks can also serve well the demands of some real-life tasks (due to their interactive nature) at the same time that help students cope with pressure. However, in group tasks teachers must ensure that all students have the same load of communicative information to produce during the task. Another assessment alternative according to the students’ opinions is one-to-one oral interviews, preferably with the teacher. Thus, a good combination of group conversations, dialogues, and interviews may compose a good repertoire of task-based language tests. Since oral tests are generally more time-consuming than written tests, it is recommended that teachers design rubrics that can be easy to use by the teacher and easy to understand by students. These rubrics must be incorporated into the task description sheet so that students can refer to them before and after accomplishing the task. It is suggested that teachers use the same rubric, if possible, for each task so that both teacher and students get familiar with them. Familiarity with the rubric will allow the teacher to maintain good control of time in order to not fall behind in the development of syllabus content. As stated throughout this article, despite the fact that task-based language teaching models have been traditionally applied in second language classrooms and special language programs (immersion, specific purpose, etc.), this experiment showed that a task-based language assessment model is possible in a foreign language classroom. As Prahbu (1987) suggested, effective learning occurs when students are fully engaged in a language task, rather than just learning about the language. This experiment also showed that a task-based language model can be implemented in beginning language classes with positive results. The implementation process may not be as smooth as that in higher usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 157 proficiency groups or that in specific purpose programs, yet a significant degree of communicative and linguistic competence can be achieved. As Willis and Willis (2007) put it, “This is one of the most valuable things we can give a learner: the confidence and willingness to have a go, even if their language resources are limited” (p. 2). Finally, this study intends to serve as the foundation or complement of further studies on task-based language learning processes in foreign language classrooms. The results obtained in this research project are satisfactory, yet it is my expectation that more teachers and researchers will carry out similar experiments that confirm not only these results but also the benefits of the task-based language teaching methodology. References Breen, M. (1989). The evaluation cycle for language learning tasks. In R. K. Johnson (Ed.), The second language curriculum. Cambridge, United Kingdom: Cambridge University Press. Breen, M. (1987). Learner contribution to task design. In C. Candlin and D. Murphy (Eds.), Language learning tasks (pp. 23-46). Englewood Cliffs, NJ: Prentice Hall International. Bygate, M., Skehan, P., & Swain, M. (2001). Researching pedagogic tasks, second language learning, teaching, and testing. Harlow, England: Longman. Candlin, C. (1987). Towards task-based language learning. In C. Candlin and D. Murphy (Eds.), Language learning tasks (pp. 5-22). Englewood Cliffs, NJ: Prentice Hall International. Common European Framework (2008). Tasks and their role in language teaching, 7, 157-167. Ellis, R. (2003). Task-based language learning and teaching. Oxford: Oxford University Press. Krashen, S. (1985). The input hypothesis: Issues and implications. Beverly Hills, California: Laredo Publishing Co., Inc. Lee, J. (2000). Tasks and communicating in language classrooms. Boston: Mc Graw- Hill. Long, M. (1985). A role for instruction in second language acquisition: Task-based language teaching. In K. Hyltenstam and M. Pieneman (Eds.), Modelling and assessing second language acquisition (pp. 83-96). San Diego, California: College-Hill Press. usinG tasks to assEss sPanish LanGuaGE LEarninG hErrEra mosquEra No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) 158 Norris, J., Brown, J., Hudson, T., & Yoshioka, J. (1998). Designing second language performance assessment. Technical report. Hawaii University, Manoa: Second Language Teaching and Curriculum Center. Nunan, D. (1989). Designing tasks for the communicative classroom. New York: Cambridge University Press. Prabhu, N.S. (1987). Second language pedagogy. Oxford, United Kingdom: Oxford University Press. Willis, D., & Willis, J. (2007). Doing task-based teaching. Oxford: Oxford University Press. Author * Leonardo Herrera Mosquera has a B.A. in Modern Languages from Universidad Surcolombiana and M.A. in TESOL from Greensboro College. He is currently working as full-time professor in the Department of Teacher Education at Universidad Surcolombiana. Email: lehemos@hotmail.com usinG tasks to assEss sPanish LanGuaGE LEarninG No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012) No. 6 (Nov. 2012)