Édgar Picón Jácome
Universidad de Antioquia, Colombia
edgar.picon@idiomas.udea.edu.co

This article was received on October 15, 2011, and accepted on May 30, 2012.

In this article I present some findings of an action research study intended to find out to what extent a teacher-student partnership in writing assessment could promote high school students' autonomy. The study was conducted in a U.S. school. Two main action strategies in the assessment process were the use of symbols as the form of feedback and the design of a rubric containing criteria negotiated with the students as the scoring method. Results showed that the students developed some autonomy reflected in three dimensions: ownership of their learning process, metacognition, and critical thinking, which positively influenced an enhancement of their writing skills in both English and Spanish. Likewise, the role of the teacher was found to be paramount to set appropriate conditions for the students' development of autonomy.

Key words: Action research, assessment for learning, learner autonomy, rubrics, summative assessment.

En este artículo presento hallazgos de una investigación-acción cuyo objetivo era averiguar en qué medida una forma alternativa de evaluación negociada promovería la autonomía de los estudiantes. El estudio se realizó en una escuela secundaria norteamericana. Las principales estrategias de acción fueron el uso de símbolos en la retroalimentación y la inclusión de criterios negociados con los estudiantes en el diseño de una rúbrica que se utilizó como instrumento de evaluación y calificación. Los resultados mostraron que los estudiantes desarrollaron su autonomía en tres dimensiones: apropiación de su proceso de aprendizaje, metacognición y pensamiento crítico, lo que influenció positivamente el desarrollo de sus habilidades de escritura tanto en inglés como en español. Asimismo se encontró que el papel del profesor es de vital importancia para establecer condiciones propicias en el desarrollo de la autonomía de los estudiantes.

Palabras clave: autonomía del estudiante, evaluación formativa, evaluación sumativa, investigación-acción, rúbricas.

In his work, Assessment of autonomy or assessment for autonomy?, Lamb (2010) elaborates on the notion of assessment that is designed to foster learner autonomy. Supporting his arguments with the work of Black and Williams (1998, 2005, cited in Lamb, 2010) and Black and Jones (2006, cited in Lamb, 2010), Lamb comes to the conclusion that "assessment for learning is designed to develop the necessary capacities for becoming an autonomous learner with a view to improving learning through better self-monitoring and self-evaluation leading to better planning" (p. 100). The author defines assessment for autonomy as "any assessment for which the first priority in its design and practice is to serve the purpose of promoting pupils' autonomy" (p. 101).

This article presents my effort within that same spirit: designing assessment procedures with the objective of enhancing students' autonomy. I highlight three aspects of the study that I consider especially significant: (1) the foreign language setting in which it took place that might provide valuable insight for schoolteachers to try out similar actions; (2) the usefulness of rubrics to help teachers make grading practices formative and provide space for them to share their power with the students; and (3) the fact of the study being inserted in a cycle of action research, which places research practices within the reach of teachers.

The assessment system in this project corresponds to a teacher-student partnership type¹ in which students participated in the creation of the scoring instrument and as active co-evaluators. Assessment had both a formative and a summative purpose within a conscious intention on my part to promote students' autonomy. In the following paragraphs, I expand the key concepts that support this assessment procedure.

Key Terms that Define Teacher-Student Partnership Assessment

Assessment and evaluation are terms sometimes used indistinctly referring to the same processes. Consequently, I find it necessary to clarify what those terms mean within the framework of my project. In doing so, I take the ideas of Williams (2003) for whom assessment designates the following four related processes: deciding what to measure, selecting or constructing appropriate measurement instruments, administering the instruments, and collecting information. Evaluation, on the other hand, designates the judgments we make about students and their progress toward achieving learning outcomes on the basis of assessment information (p. 297). Brown (2004) expands assessment definition asserting that it is a continuous process that takes place either on a formal or an informal basis.

Assessment and evaluation can be classified according to the focus of power. For instance, teacher-student partnership, a concept developed by Bratcher and Ryan (2004), is a type of evaluation in which both teachers and students work together. Some key words that describe this approach to grading are input (from both sides), negotiation, and flexibility. Power is not concentrated on the teacher but shared with the students and there is a continuous combination of different student and teacher roles in every step of the process. Bratcher and Ryan assert that this type of evaluation has the advantages of students "investing in grades in which they feel they have had input" (p. 102).

Assessment can also be classified according to its purpose -when it is administered and how its results are used- as either formative or summative. Formative assessment aims at measuring achievement within the process and helping students to improve their skills. Contrariwise, summative assessment measures results at the end of a process mostly in order to make decisions (Angulo-Delgado, 2002; Arias, Estrada, Areiza & Restrepo, 2009; Bachman & Palmer, 1996; Brown, 2004; Ekbatani, 2000; Himmel-Köning, Olivares-Zamorano, & Zabalza-Noaim, 2000; and Lippman, 2003). This study takes place within a context in which grading plays an important role and therefore assessment has a summative purpose. However, the action strategies applied had the intention to make it formative by providing space for feedback -among other strategies- which is expanded in the following section.

Feedback

Feedback is inherent to formative assessment. In fact, Goodrich-Andrade and Boulay (2003) -citing Cooper and Odell (1999)- define assessment as "ongoing feedback that supports learning" and stress the need of providing students time for reflection upon and self-assessment of their pieces of writing before they submit a final draft (p. 21). Arias et al. (2009), in addition, assert that there must be a continuous and systematic process of feedback for formative assessment to be successful.

One of the forms of feedback that I used in this study is in agreement with Rutherford's arguments in favor of the teaching of grammar rules. Rutherford (as cited in Edlund, 2003, p. 369) argues that adult learners go into a process of comparison between the two grammatical systems in which they make and test theories about how L2 works. The process of producing such theories can be facilitated by what he calls "grammatical conscious raising" or C-R. C-R is the supplement of data needed during the theory testing occurring in the L2 learner's mind. Edlund (2003) thus points out that this theory justifies the practice of selective marking of errors, which was applied in this study as part of the action strategies.

The other form of feedback was the use of analytic rubrics for self-evaluation. Rubrics have been found to be useful to provide both formative and summative feedback in a systematic and effective manner (see O'Malley & Valdez-Pierce, 1996; Mertler, 2001; Moskal, 2000; and Stevens & Levi, 2005). While the use of symbols for self-correction emphasized syntax and vocabulary, the rubrics included aspects of the discourse component that complemented the linguistic construct evaluated². In the following section, I expand the definition of rubric.

Rubrics

Mansoor and Grant (2002) define a rubric as "a scoring device that specifies performance expectations and the various levels at which learners can perform a particular skill" (p. 33). This is the concept of rubric that applies to the scoring method employed in the study, and the same that authors such as O'Malley and Valdez-Pierce (1996), Moskal (2000), Mertler (2001), and Stevens and Levi (2005) are in concordance with.

Rubrics are pertinent for criterion reference assessment since they provide the space for assessment criteria to be explicitly stated (Bachman & Palmer, 1996; Brown, 2004; Genesee & Upshur, 1996; Himmel-Köning et al., 2000). Likewise, they are coherent for the scoring of constructed-response assessments (Brown & Hudson, 1998) such as the short compositions that the students produced for this study. Rubrics need to be closely connected to the task that they will score and the task should clearly state the specific and detailed information that the students will need in order to complete it successfully (see O'Malley & Valdez-Pierce, 1996). Rubrics can be easily accessed and downloaded from internet sites. However, following the ideas of Hewitt (1995), I decided to design the rubrics along with my students in order to facilitate discussion and reflection about the criteria.

In the words of Black and Jones (2006), "an assessment activity can help learning if it provides information to be used as feedback" (cited in Lamb, 2010, p. 100). Teacher-student partnership is consequently assessment for learning. Following Lamb's (2010) definition of assessment for autonomy, we can finally assert that the assessment system applied in this research project had the characteristics of such assessment practice. Next, I elaborate on the definition of learner autonomy.

In the words of Benson (2010), when we talk about autonomy in language learning, we usually "refer more to a certain kind of relationship between the student and the learning process." (p. 79). The author asserts that the term that best describes this relationship is control. Following this order of ideas, Benson offers a framework to measure learner autonomy. This framework is represented by three poles of attraction among which various degrees of control over learning could be determined: those poles are student control, other control and no control (p. 80). Learner autonomy could thus, to some extent, be evaluated in relation to a student level of control, at a certain point in time, over dimensions of the learning process such as "location, timing, pace, sequencing and content of learning" (p. 79). From that perspective, evaluation being another dimension of the learning process, one might establish students' development of autonomy in terms of the student level of control over it at different points in time.

Dimensions of Learner Autonomy

Previously, in his well known work The Philosophy and Politics of Learner Autonomy, Benson defines three dimensions of autonomy -technical, psychological, and political. The technical dimension concerns the techniques and strategies that help students to become owners of their learning process i.e. individuals with the capacity to manage their own learning. In order to facilitate its development, it is paramount to promote self-directed learning, which includes providing students situations for them to learn how to learn (Benson, 1997).

Concerning the psychological dimension, Benson considers that it involves the development of traits in the individuals that leads them to become more responsible, develop critical thinking, and take control over their learning process. Learners are the ones who construct knowledge starting from their social interaction and continual self-evaluation that should lead to self-awareness.

With regard to the political dimension, Benson asserts that it relates to the learners' ability to deal with power issues within the teaching-learning process. Benson highlights that whether the teacher takes full control of the power within the classroom or whether s/he decides to share it with the students is a political decision that affects learning completely. In the same way, Benson and Voller (1997) affirm that learner autonomy "can be thought of in terms of (...) redistribution of power among participants in the social process [of education]" (p. 2); hence, the development of a more political dimension of learner autonomy could be facilitated by teaching methodologies in which students have the opportunity to participate in decision making.

The role of the Teacher

Many authors have emphasized the role of teachers in the promotion of learner autonomy (see Benson, 1997; Ellis, 2000; Lamb, 2010; Little, 1995; O'Malley & Valdez-Pierce, 1996; Voller, 1997; and Wenden, 1991). In this sense, the teacher is a facilitator, counselor or guide with a supportive attitude towards the learner and within a learner-centered environment; a teacher is willing to release some power over the students in behalf of their development as independent, able learners. Furthermore, they have pointed out the possibility to help students develop autonomy by teaching them strategies to learn the language, rather than transmitting the language, and fostering self-reflection and critical thinking.

Following this rationale, Wenden (1991) examines the features of autonomous learners, shows how those characteristics are linked to learning strategies, and proposes activities to teach those learners. In her analysis, the author uses the typologies of learning strategies defined by Chamot (1987). Wenden groups learning strategies, according to their function in the learning process, as cognitive and self-management.

Self-management, which corresponds to O'Malley and Chamot's (1990) metacognitive strategies, include three main functions: planning, monitoring, and evaluating. Planning has to do with specifying matters such as time, place, individuals, resources, forms, and reasons to carry out an activity or to state a task leading to learning the language. Monitoring has to do with constantly identifying failures in the act of communicating while the communication is taking place. Finally, evaluating has to do with reflecting on the development of the strategy planned and its pertinence in terms of learning.

Besides metacognitive strategies, O'Malley and Chamot's (1990) typology point out two more main types: cognitive strategies and social/ affective strategies. While cognitive strategies are "more directly related to individual learning tasks and entail manipulation or transformation of the learning materials" (p. 8), social/affective strategies are related to cooperative processes of learning and the control of affective matters that affect the language learning process.

A failure identified in this study is that learning strategies were not explicitly taught; they were identified in the analysis, however. In my belief, teachers have a special responsibility to help students develop autonomy in a more political dimension. Training students to self-evaluate against clear criteria, and giving them the opportunity to act as co-evaluators, is a way both to help them foster metacognition in L1 and L2-English and Spanish in this study- and to develop skills in order to fight for their rights; a movement of evaluation practices towards fairness and democracy (Shohamy, 2001).

This study is framed within a cycle of action research. Action research has been found to be especially appropriate for educational improvement (see Altrichter, Posch & Someck, 1993; Selener, 1997; Burns, 1999). One of the goals of action research is to involve teachers in reflection upon and within their practice so that they (1) become aware of the possible factors that might constitute a particular question/problem encountered in their day-by-day teaching lives, (2) better understand those factors, and (3) plan and carry out strategies in order to find answers to and/or solve that question/problem. Burns (1999) found that doing action research enabled teachers "to engage more closely with their classroom practice as well as to explore the realities they faced in the process of curriculum change" (p. 14). In addition, it produced "personal and professional growth" and increased teachers' "self-awareness and personal insight" (p. 15).

The Starting Point

My starting point for this project came from my experience teaching Spanish as a foreign language in high school in the U.S. English was the mother tongue for the majority of the students³ and Spanish classes took place two hours daily. High school students needed to complete at least one year of a foreign language in order to graduate and Spanish was the favorite one due to the Hispanic population growth that was turning some environments bilingual in the U.S.

At the time of this proposal, I was teaching eleventh and twelfth graders, and the group with which I systematized the experience was composed of 19 students, eight girls and 11 boys, whose ages ranged from 15 to 18. Relating to their placement in the school, there were 10 seniors and nine juniors. Two of the students dropped the course.

I had found that many of my Spanish students did not keep track of their notes and then depended on me or other students to choose the vocabulary needed for their writing tasks and to correct their grammar errors. Some students seemed not to have learned the mechanics in previous classes, made mistakes that they did not know how to correct, and felt unmotivated towards the writing task.

Many students seemed not to care about the whole process and to lack clear objectives and/or reasons for studying the language; although speaking Spanish was considered an advantage for potential jobs in the future, most students expected to learn it without much effort. The majority relied on English to communicate in class. As a consequence of these conditions, some students copied from partners and did not even worry about learning while others translated whole papers using computer software without knowing what they had written. The writing activity ended up being of little value for such students. I believed that the previous situation was directly related to students' lack of ownership i.e. lack of autonomy.

I had already applied some of the assessment procedures that I incorporated into the project with the feeling that they helped me to become more successful in my teaching. Nonetheless, I had not taken into account my students in the development of evaluation criteria, neither had I given self-evaluation much importance in their final grades as I did in this course.

Action Plan

The main action that I applied in order to cope with the situation described in the previous section was the implementation of a teacher-student partnership form of assessment. Two main strategies in this assessment process were (a) the implementation of self-correction of errors by using symbols as a form of feedback (see Appendix A) and (b) the design of a rubric containing criteria negotiated with my students in order to self-evaluate and grade some of their compositions. The plan would be applied through different steps including training students both to self-correct their errors, and self-evaluate against the criteria stated on the rubric. The plan was intended to fit within my normal teaching activities.

Data Collection

The data collection included my personal journal; a survey of the students; students' products, namely two compositions plus early and late writing samples of theirs and the scoring rubrics used to evaluate their performance. For the purpose of this article, I took into account the findings that emerged from the analysis of my journal and the survey itself in addition to inferences made from the students' progress based on the analysis of their scores.

In my personal journal, I described the implementation of the actions that I had planned and reflected on every step of the process. I registered in it my interpretations of the outcomes that emerged from the analysis of the survey and the comparison of self and teacher assessment. I also recorded my personal opinion in the journal of the students' performance and progress through the implementation of the strategies, which gave me the possibility to triangulate my perception with my students' and to keep track of the chronology of the events.

Twelve students responded to a survey given at the end of the course. It was a brief survey composed of four multiple choice questions and an open-ended one aimed to find out the students' sense of whether the procedures used had helped them or not in terms of developing strategies and/or attitudes towards their process of learning how to write in the foreign language.

The students wrote, self-corrected, and self-evaluated two compositions for the purpose of this research: Mi Escuela and Mi Familia (see Appendix B). In total, thirteen students' self-evaluation forms were included in the analysis of the first composition and eleven in the second one.

Data Analysis

In order to analyze qualitative data, such as narrative and descriptive events, personal reflections, and open-ended questions, I carried out inductive-deductive analyses following the steps suggested by Burns (1999): assembling, coding, comparing and building interpretations (pp. 156-160).

For the analysis of quantitative data, such as the comparison between the student's and the teacher's evaluation forms, and the outcomes of the survey, (1) I created charts and tables comparing the results; (2) I described those results and tried an initial interpretation through a reflection exercise in which I tried at connecting them to the main topics of the project; and (3) I carried out inductive deductive analysis of those descriptions and reflections i.e. I categorized them too.

I invited my students to participate in the process by giving me feedback about the different strategies and about the process in general. At the same time, I asked my students for permission to use their pieces of writing and signed a compromise letter committing myself to guarding their anonymity in order to be consistent with the ethical principles of educational research (Pring, 2004; and Burns, 1999). Finally, I asked a colleague to act as a critical friend in order to enhance validity of the data analysis (Altrichter et al., 1993).

In this section, I discuss the findings of this study that intended to analyze to what extent a teacher-student partnership on writing assessment could promote students' autonomy. Analysis of the data showed that some development of learner autonomy resulted from the interaction of the teacher's role and the actions taken in this research project. Learner autonomy was thus reflected in three student features: gaining ownership of their learning process, developing metacognition, and developing critical thinking.

Gaining Ownership

In the framework of this discussion, gaining ownership refers to students' actions, behaviors, or attitudes that showed their movement towards a more autonomous dimension of learning. Students showed that they had gained ownership by expressing or showing independence, showing commitment and responsibility to do the learning activities proposed, participating in decision making, and expressing their having felt part of the development of the assessment system.

To start off, some of the students' responses to the survey confirmed a positive attitude towards the use of social and cognitive strategies after having been involved in the project: on the one hand, 90% of them acknowledged that they were more likely to ask for help instead of copying their partner's work; and on the other hand, 60% of them expressed that they were more likely to use their notes (see Figure 1). I, myself, corroborated such behavior during our visits to the lab and in other opportunities in class:

...they would sit next to their partners and ask them for help... I thought that was part of the plan; they should be willing to ask for help instead of just copying. ...Also, most of the students were using their notes and dictionaries. (Journal, p. 15)

By the same token, 50% of the students agreed that the activity of self-correcting their own errors helped them to become more independent and it is particularly significant that one of them expressed that she was more likely to "do better herself"⁴ after the experience (see Figure 2). Consequently, since the use of learning strategies implied students' progress in terms of independence and positive attitudes towards learning, it also evidenced some development of ownership of their learning process, which has been associated with learner autonomy (Lamb, 2010; Benson, 1997; Little, 1995; Wenden, 1991).

It seems that the assessment process enhanced most students' commitment towards the development of the activities completed in class. This could be observed in their changing attitudes, keeping on-task, and expressing pride in their work, as I recorded in my journal.

Generally speaking, I noticed that most of the students would be concentrated in their job...With the exception of few students, I felt that the activity had engaged them. ...Another positive aspect I noticed was that most of them looked proud of their work. They would decorate their final papers and use fancy font. (Journal, pp. 15 & 16)

Most of the students who showed little motivation and commitment towards class activities in the beginning of the semester gradually changed their attitude. Benson (1997) points out that "constructivist approaches to language learning tend to support 'psychological' versions of learner autonomy that focus on the learner's behavior, attitudes and personality" (p. 23). He goes on to assert that those versions "can be seen as promoting qualities in individual language learners that will be of value in the process of independent language use" (p. 29).

On the other hand, a more observable behavior leading to the development of ownership was students' actual participation in decision making. Benson and Voller (1997) state that autonomy can be understood as a right of learners to direct their learning process; hence, active participation in deciding criteria against which they were going to assess themselves was evidence of students learning how to take control of that part of their learning that involves evaluating their achievement.

It was very gratifying to see that most of them actually discussed the criteria and gave me feedback in order to design the rubric. As I recorded in my journal, the majority participated with their comments, which were very valuable in the design of the assessment instrument (Journal, pp. 6-11). Likewise, the results from the survey showed that, as a consequence of participating in the design of the rubric, 70% of the students felt part of the grading process and 48% felt that they had been taken into account in decisions that affected their performance (see Figure 3).

The act of participating in the establishment of something as determinant of power as the grading criteria in a language course implied a movement towards a more political dimension of autonomy in my students. As Little (1995) warns:

In formal educational contexts learners do not automatically accept responsibility for their learning -teachers must help them to do so; and they will not necessarily find it easy to reflect critically on their learning process-teachers must first provide them with appropriate tools and with opportunities to practice using them. (pp. 176-177)

Because it is a determinant factor for students to become better able to self-direct their learning within the psychological dimension of autonomy, in the following section I give special emphasis to the students' use of metacognitive learning strategies.

Developing Metacognition

Metacognition, which relates to mental processes that involve reflecting, comprehending, interpreting, reexamining, planning, monitoring, self-evaluating, and, generally speaking, expressing self-awareness of learning, is very important in the development of autonomy because it enables human beings to self-manage their learning process. The three main pieces of evidence found in the data analysis as development of metacognition in the students were their development of metacognitive learning strategies, namely, planning, self-monitoring, and self-evaluating. It is important to clarify that these processes were carried out in English, which evidenced the role of L1 in scaffolding the development of L2⁵.

The evaluation criteria stated in the rubric -along with the detailed guidelines for the development of the task- demonstrated to have been helpful for students to have a clearer idea of what they had to do in order to be successful in the writing assignment. Hence, the students' increased awareness about the general process of the task must have influenced the fact that most of them planned better and thus improved their grades for the second evaluation event. Table 1 shows comparison of the students' final grades for the two compositions.

The responses for question 2 of the survey also showed that 85% of the students found participating in the design of the rubric useful to better understand what was expected from them (see Figure 3). According to some of the students' actual words,

...designing a rubric with the teacher:

Help me understand what I need to include in my writings

It does make me know what to expect, what is going to be required and that it is part of the grading process

I will know what is expected to be on an assignment that will give me a 100.

An outstanding positive result of students' participating in the design of the rubric was their awareness of the features of a composition that would meet the standards for a good grade. It must have been determinant for their planning since it helped them to organize their ideas and apply their language knowledge while writing. In our case, most of the students (85%) identified the development of planning skills in terms of organization of ideas around a given topic as the second clearest improvement that resulted from their participation in the rubric design (see Figure 4). Students' own words picture their point of view:

It did help me organize on a topic because it gave me ideas what I was going to write about and it helped me with Spanish grammar... It also helped me in my ability to write a composition more carefully and it gave me ideas. (Questions 3 & 5, survey)

In a broader dimension of autonomy, planning would imply organizing a more complex set of details such as choosing times, resources, and places most appropriate for learning. However, developing planning in the restricted sense found in this analysis is surely a valuable evidence of the development of metacognition that could eventually lead the students through the path of learning how to learn.

Another self-management strategy (Wenden, 1991), self-monitoring, seems to have been fostered by the use of symbols as the form chosen to give feedback to the students' pieces of writing. In their responses for question 1 of the survey, 85% of the students identified an increased awareness of grammar errors as a result of this formative assessment strategy (see Figure 2). Students acknowledged that the activity actually helped them to better understand their mistakes and correct them, which is proof of the presence of instances of comparing and making hypotheses about how the foreign language works in terms of grammar (Edlund, 2003). Some of the students' responses to the survey evidence their development of self-monitoring as a result of self-correcting their errors based on codes provided by the teacher:

It helps you see what you did wrong and not do it again

Helped me to understand my own mistakes so I could recognize them later on

By seeing the errors that I made, it helped me to prevent me from making the same mistake. (Questions 1, 3 and 5; survey)

The key aspect in these pieces of evidence to identify students' self-monitoring was the fact that they emphasized understanding, seeing, correcting, fixing and preventing the making of mistakes as the result of the activity.

Wenden (1991) marks the difference between monitoring and evaluating, emphasizing that "in contrast with monitoring,... when learners evaluate, they consider the outcome of a particular attempt to learn or use a strategy; the focus is on the result and the means by which it was achieved" (p. 28). Some students' responses that evidence self-evaluating pointed at different aspects of the FL learning process. When a student stated, "If I use my notes, it will help me to understand better and study," he evaluated the usefulness of a strategy. Or when another student stated that the project had been "Helpful in her understanding of the Spanish language and grammer" [sic], she was referring to one specific component of her communicative competence. On the other hand, when this other student wrote, "I think doing all this will improve my English and Spanish language," he was reflecting on the usefulness of having been part of the project in general. Other students' responses that showed self-evaluation are

It helped me with my writing completely

Has improved my Spanish skills and knowledge

A good learning experience, because it really helped me learn how to correct my problems. (Questions 1 and 5, survey)

In the following paragraphs, I will discuss the appearance of a more political dimension of autonomy found in the analysis: gaining critical thinking.

Gaining Critical Thinking

Gaining critical thinking within the framework of this experience refers to events in which the students make honest judgments about their own language performance in a task based on previously agreed criteria; they also discuss and/or question decisions and reach agreements based on arguments as well as make changes towards a more responsible and successful attitude if needed. Critical thinking evidences the presence of psychological and political traits of autonomy that should eventually be transferred to other situations.

The psychological traits involve values such as responsibility and honesty while the political traits imply learning how to take a stand and support one's point of view. In its political dimension "learner autonomy represents recognition of the rights of learners within educational systems" (Benson, 1997, p. 29). Consequently, "a considerably expanded notion of the political... would embrace issues such as... roles and relationships in the classroom and outside, kinds of learning tasks, and the content of the language that is learned" (p. 32). Following, I present evidence of an opportunity for a student to discuss his grade based on arguments, which reflect his development of critical thinking:

When I returned the papers to the students, one of them was not happy with his grade. So I discussed the disagreement with him. It was a student who was very strong, regarding speaking skills and vocabulary, but had some writing weaknesses...I showed him that he had had a lot of spelling mistakes and emphasized that I had given him the opportunity to correct before I graded the final paper. He had refused to correct his mistakes and now seemed to feel disappointed about the grade. It was not difficult to convince him; he actually smiled at the fact that I was right, and told me I could keep the grade like that. (Journal, p. 13)

Besides an opportunity for the students to develop some critical thinking, this was an example of the usefulness of the rubric for me: the student could not deny that I was right because the criteria were clear and he had taken part in their establishment. However, the most significant evidence of development of autonomy for this kid was his change of attitude for the rest of the semester; he became more responsible and careful in the development of the writing tasks. This could thus show that his growth in self-criticism influenced his gaining ownership.

More evidence for development of critical thinking can be inferred comparing the number of discrepancies between teacher and students' grades during the first and second instances of self-assessment. The fact that this number decreased significantly in the second instance proved that students' abilities to self-evaluate their work improved with experience and that students must have become more self-critical as they gained expertise evaluating their work (see Table 2).

The Teacher: Factors that Determined my Role

Given that this study presents teaching strategies applied within a conscious effort of the teacher to promote autonomy in his students, the teacher's role in the process is important to be discussed. The main themes that emerged in the data analysis regarding my role as a teacher in the promotion of learner autonomy were empowering, fostering critical thinking and guiding.

Empowering was evident in different forms, namely, strengthening students' independence by training them to self-correct and self-evaluate as well as teaching them the use of cognitive strategies such as the use of the dictionary and memory strategies, and encouraging their participation in decision taking. My beliefs in regard to assessment as a democratic formative process in which learners should play a starring role were decisive for the traits of learner autonomy identified in the analysis to take place. My teaching methodology might as well have facilitated empowering and thus autonomy: the analysis of the data showed that I used a learner-centered approach to teaching characterized by group work, the use of technology, and differentiated instruction, which promoted the development of social learning strategies, facilitated students' independence and added to their technical dimension of autonomy.

My role in the development of students' critical thinking was evident in two main instances: (a) the promotion of discussion among them to select the criteria for the rubric, and with me to discuss grades based on those criteria; and (b) the facilitation of students' self-assessment itself. Part of the impact of students' self-assessment could be identified as the appearance of self-criticism and this was evident in some of them changing negative attitudes and committing themselves to the writing tasks. Some aspects of my personality probably facilitated this process. My own self-criticism provided space for reflection and allowed the benefit of the doubt about me being always right, which was decisive to acknowledge students' rights and prevented me from taking authoritarian decisions. Likewise, my flexibility might have helped in decreasing students' anxiety at the time of discussing grades, making it easier for me to approach them and question their behavior. Both aspects support Voller's (1997) statement that "teachers need to reflect critically not only upon how they act during a learning event, but also upon their underlying attitudes and beliefs about the nature of language and the nature of learning" (p. 112).

My role as a guide and technical support was evidenced during the development of the key action strategies of the project. At the time of defining evaluation criteria, I negotiated with the students, on the one hand, and, on the other, either used my expertise or linked theory and practice in order to make an informed decision. Likewise, I guided students to self-correct by using symbols and using the rubric for self-evaluation. Again, both languages had their role in this process: English was the one used to lead reflections about meta-linguistic aspects and Spanish the goal in terms of communicative competence.

Teacher-student partnership assessment proved to be a valid strategy to promote learner autonomy. The findings showed that three dimensions of it developed in the students who took part in this study: ownership of their learning process, metacognition and critical thinking, which were found to interrelate producing better conditions for learning. Regarding ownership, the students showed some independence from the teacher and some sense of responsibility, both of which were evident during in-class activities that required the students to be involved and committed. This development of ownership seems to have been positively affected by a movement observed in some students towards a more self-critical thinking as evidenced by their recognition of and effort to cope with negative attitudes. Alternatively, a more responsible and committed attitude surely helped to prepare the terrain for more independent learning, supported by the two main action strategies taken in the project. In that sense, the students' development of some planning, self-monitoring, and self- evaluating skills seems to have positively affected their writing achievement, and showed an increment on their part of both their technical and psychological dimensions of autonomy.

Critical thinking was equally reflected in the fact that students' self-evaluation was more accurate during the second evaluation event, which could be explained based on students' better understanding of evaluation criteria as they gained experience using the rubric. As a result, the data showed that some participants were developing the ability to support their self-evaluation by acting as trained co-evaluators. From this perspective, students' development of critical thinking seems to have a direct correlation with the formulation and discussion of clear evaluation criteria, which preceded the design of the rubric. By the same token, this finding evidences a movement of the students towards the development of a more political dimension of autonomy.

This experience of teacher and students becoming partners in the process of assessing for learning has equally proved that the teacher's role is to some extent a dependent factor of learner autonomy in the school context. The main features that emerged in the analysis proving such statement were my role as a guide and technical support, which reflected my moving towards a more learner-centered teaching approach; and my beliefs in more democratic forms of assessment, which provided space for students' participation, fostering their critical thinking and empowering them. Using Benson's framework, one could say that, in terms of evaluation, learner autonomy moved towards the student-control pole with the support of the teacher.

In spite of these positive results, I am aware that learner autonomy has a multidimensional nature and needs to be analyzed from different perspectives. My analysis is thus limited by at least two aspects: (1) that only one dimension of learning was taken into account i.e. evaluation; and (2) that learning was only observed inside the classroom within a somehow teacher-centered environment. I therefore present these conclusions with an awareness of such limitations. This action research project was implemented upon a sample of convenience and conclusions are subjected to generalization only to the extent to which the reader identifies similarities in his/her teaching-learning context.

Although teacher-student partnership assessment has been confirmed to be a significant means to promote learner autonomy, a political dimension needs to be addressed at a more critical level; promoting student reflection about social issues that affect them could be an objective in teachers' planning that would help the latter to achieve such a goal.

Regarding the assessment construct, teachers must keep updated with the models that support their teaching approach in order for their assessment criteria to be clear and valid. Likewise, content and performance standards need to be established among the language teachers of the school in order for them to be able to design assessment procedures that could more accurately evaluate students' level of proficiency.

This study offers valuable insight for language teachers in high school about the possibilities of democratic assessment practices to support language learning. By the same token, it provides a view of the conditions under which their colleagues teach and allow themselves to compare and find similarities. EFL/ESL teachers will surely identify challenges in this article similar to the ones they face in their daily practices. This case is consequently evidence that it is possible to change things for the better in such contexts.

Finally, I have realized the benefits of carrying out action research via my better understanding of the situation and my learning about the topics that framed the project. Furthermore, I became more conscious of my role and responsibility in the development of my students' autonomy at the time that I developed mine.

¹ Although Bratcher and Ryan focus on evaluation and grading for their classifications, I consider that the whole process in this study actually corresponds to my definition of assessment; consequently, I have called it Teacher-Student Partnership Assessment.

² The Communicative Competence was the theoretical construct evaluated.

³ One of the students had Hispanic roots and there were others with immigrant parents. The school itself had many students who were Hispano immigrants or had Hispanic roots.

⁴ Emphasis mine.

⁵ Evaluation criteria and symbols were written in English. Teaching took place in both languages. Compositions were written in Spanish.

Édgar Picón Jácome holds an MA in TESOL from Greensboro College in Greensboro, USA. He is an assistant professor at the School of   Languages of   Universidad de Antioquia, Colombia, and belongs to the GIAE Research Group. His interests in research include teacher and learner autonomy, and evaluation.