A bridge too far? Dynamic Assessment and a dialectical view of the gap between language assessment and learning

Julkaistu: 8. marraskuuta 2023 | Kirjoittanut: Matthew E. Poehner

Introduction 

The divergence of language assessment from language teaching and learning is certainly familiar to classroom teachers, who have to shift from one to the other, adopting different materials and methods and engaging with their students in very different ways. The divide is also frequently found among researchers. Indeed, the second language (L2) field, like education more generally, includes professional organizations with regular conferences, research journals, and even academic departments or faculties at universities that often signal a commitment to either assessment (testing, evaluation, measurement) or teaching and learning (instruction, pedagogy, curriculum). Add to this the highly specialized nature of much academic research, and it becomes perhaps less surprising that scholars, educators, learners, and the public perceive a gap between assessment and learning. Indeed, the proliferation of terms that add various modifiers to the word ‘assessment’ (e.g., formative assessment, assessment-for-learning, assessment-as-learning, learning-oriented assessment, etc.) further evidence acceptance that such a gap exists and concern that it does not serve the needs of teachers and learners. 

Against this backdrop, one might be forgiven for responding to the perennial call to narrow the assessment-learning gap through some innovation by asking, is that a bridge too far? Is closing, bridging, or narrowing the gap simply too ambitious an undertaking? In this paper, I begin with a different question: What if the gap need not exist at all? This is not idle speculation but a real possibility, albeit one that depends upon the theoretical perspective one adopts. The perspective to which I am referring is Sociocultural Theory (SCT), which originated in the writings of Russian psychologist L. S. Vygotsky and his collaborators. SCT suggests to us a very different starting point when considering assessment and learning, specifically that the two actually only make sense when understood in relation to one another. This idea of relation, which I refer to as dialectical in this paper, is important because it does not imply that assessment is the same as teaching and learning (a relation of identity) and neither does it propose a dichotomy or relation of opposites. In what follows, I briefly explain this dialectical relation by discussing Vygotsky’s well-known conceptualization of the Zone of Proximal Development, or ZPD. I then turn to an assessment framework that this dialectical way of thinking has given rise to, namely Dynamic Assessment (DA). DA examples are offered to illustrate this framework and how it may guide us over the bridge. 

The essential relation of assessment to teaching and learning: The ZPD 

Dialectics is a way of thinking that reframes many perceptions of the world that we take for granted by focusing our attention on phenomena not as discrete, separated from one other but as existing in sets of relations. Categories that are often viewed as opposites, such as day and night, nature and nurture, internal and external, are instead seen as different features of a broader phenomenon, which can only be fully understood by taking account of both these parts. Moreover, these elements can only be understood when they are viewed as part of these relations. Our conception of daytime only makes sense in relation to night and as part of a larger analytic unit, the 24-hour cycle we refer to as day. 

Vygotsky applied dialectical thinking in both his scholarly research and practical activity as a psychologist and educator. In fact, dialectics was so fundamental to his work, that his scholarship, including theoretical advances, and his practical work with teachers and children were mutually dependent on one another. The ZPD appears in Vygotsky’s writings as an essential feature of his account of how our cognitive, emotional, and intellectual abilities develop. Specifically, it is the transformation of our natural psychological processes into ones that are mediated through our social interactions and our use of available cultural resources, including language, numbers, concepts, and the various other affordances available to us. It is in this regard that Vygotsky (2012) discussed the ZPD as helpful to understanding how people function individually and how they function when various forms of mediation that are available. According to Vygotsky, observing a person complete tasks independently, as occurs in most forms of assessment, is helpful in so far as it reveals development that has been completed at the present point in time. Development in this model refers to the internalization of forms of mediation (e.g., planning, reasoning, reflecting, and the use of resources such as language and conceptual knowledge in particular fields of study) that learners first encounter during interaction with others. When working independently, we mediate ourselves. For Vygotsky, however, this reveals only a partial view of our abilities because it does not illuminate those abilities that have begun to form but have not yet fully developed. Vygotsky (2012) used the metaphor of the ‘buds’ of development versus those that have fully flowered and yielded ‘fruit.’ He argued that the gardener must take account of not only the fruits (i.e., abilities that have fully developed) but also the buds (i.e., the abilities that are still emerging). These latter abilities, the ‘buds’, constitute the ZPD, the abilities that will soon ripen. Importantly, the way to include these abilities in an assessment is to look not only at a person’s independent functioning but also at how they respond when their performance breaks down and another person intervenes to mediate their performance. As Vygotsky explained, this intervention may take the form of offering a model or example for learners to follow, providing leading questions, prompts, or feedback, and beginning to complete the task and asking the learner to take over. 

Two observations are in order. The first is that another person such as an assessor or teacher intervening during the assessment to provide mediation to learners constitutes a special form of teaching as part of the assessment procedure. To reiterate, the purpose of this intervention is not to help the learner achieve a better score on the assessment but to identify the ZPD. This is important because people may differ with regard to how responsive they are to mediation during the assessment; some may improve their performance only slightly, for example, or some may require very little intervention, such as simple reminders. This leads to the second observation: identification of the ZPD for individual learners can help teachers to better align their instruction to learner needs. Together, these two observations help us to understand how the ZPD binds assessment with teaching and learning in an essential relation. 

L2 DA: Identifying the ZPD to promote learner L2 development 

Since Vygotsky’s time, various assessment procedures have been devised that include the ZPD within their scope, and collectively these are referred to as ‘dynamic assessment’. The term dynamic assessment signals that the procedures themselves are dynamic (i.e., varied as differing forms and amounts of mediation may be required by individual learners) and that learner abilities are also understood as dynamic (i.e., they are ripening or emerging and appear through responsiveness to mediation during the assessment). Today, a wide range of approaches to DA can be found, reflecting the diversity of contexts in which researchers work (i.e., formal and informal learning environments around the world), their focal populations (e.g., individuals with special needs, young children, gifted learners), and the specific abilities in question (e.g., basic cognitive functions, early literacy, mathematical reasoning). Haywood and Lidz (2006) offer a useful overview of this work. As they explain, these approaches to DA share a view that the inclusion of mediation during assessment expands how we can understand their needs and abilities, and this information is essential for orienting instruction that can be maximally beneficial to their ongoing development. 

Use of DA with L2 learners began in the early 2000s. Early research sought to elaborate the process of mediation and the interplay between mediation and learner responsiveness could be systematically interpreted to gain insights into emerging L2 abilities (Poehner, 2008). This work identified important principles, including that mediation should begin in an implicit way, such as encouraging learners to re-attempt a task or suggesting that they check their response, and then it may become more explicit only as required by the learner. DA researchers reported sets of mediating moves, arranged from most implicit to most explicit, that were used in their procedures. The importance of this implicit-to-explicit logic is that by offering learners a graded series of prompts, leading questions, models, and feedback, it becomes possible to ascertain the degree of mediation required for learners to identify difficulties and overcome them (e.g., did a learner require only a reminder or leading question or was it necessary to provide explanations and examples?). This, in turn, provides essential diagnostic information concerning how near or far a learner is from more independent functioning. 

Following this initial work, DA was extended from one-to-one procedures to whole-classroom environments (e.g., Poehner, Infante, & Takamiya, 2018) with an aim of helping teachers to track learner development over time and to use this information to better align instruction to the needs of individuals and the class as a whole. In that particular study, the teacher placed learners into groups according to common areas of difficulty and provided them with exercises that required them to work together to propose solutions. Their cooperation offered opportunities for them to attempt to support one another’s development, and the teacher was also present to employ a dynamic procedure of offering mediation (prompts and leading questions) to the groups. At the same time, other researchers began investigating implementations of DA in more formal assessment contexts (i.e., testing) and for purposes that include, among others, placing learners at an appropriate level of language instruction (Antón, 2009) or determining learner preparedness for advancing to the next level of study or program completion (Levi, 2017). In their timeline of L2 DA research, Poehner and Wang (2021) identify studies with learners from young children through adulthood, at beginning levels of proficiency to advanced, studying a wide range of languages, and targeting every communicative modality. 

Recent Innovations in L2 DA 

DA continues to be employed with L2 learners at an ever-increasing rate, resulting in a rapidly expanding research literature. As just one example of new directions being explored, Randall and Urbanski (2023) report how technology may be leveraged to conduct DA with potentially large numbers of L2 learners. The authors conducted a study evaluating the effectiveness of a computerized DA (C-DA) procedure for understanding the knowledge of and control over features of L2 Spanish grammar among U.S. university learners. A particularly innovative aspect of Randall and Urbanski’s computerized program is that it offers greater potential for mediation to be responsive to learner needs than has been the case in previous C-DA approaches. As the authors explain, C-DA requires that mediation be prepared in advance and included with assessment items and tasks in the computerized procedure. Recognizing that this limits the flexibility that a human mediator would have in pursuing different kinds of problems individual learners might experience, Randall and Urbanski designed their C-DA in such a way that different sequences of mediation for a given assessment item become available depending on the nature of a learner’s difficulty. 

The authors demonstrate the mediational process with an example in which learners are presented a sentence in Spanish that they must complete by selecting the appropriate verb form (Randall & Urbanski, 2023, p. 55). The sentence Tú______ bien después de hacer ejercicio (You ____ well after exercising) appears and students must complete it with the verb dormir (to sleep). The sentence is accompanied by a selection of options, including the correct form, dormirás, which is inflected for the second person singular form (you will sleep), and multiple distractors. If the student selects the correct form, the message Well done! appears and the test continues. If, however, an incorrect response is selected, then the mediation sequence begins. Learners are first offered a general prompt, such as Check your answer, and are encouraged to try again. If the learner selects the same or a different incorrect option, then mediation is offered that responds to the nature of the error. For instance, if a learner selects the option dormirá, which is the third person singular form (she/he will sleep), then prompts direct student attention to the subject of the sentence and the need to identify the corresponding verb form. If a student selects the option duermirás, which follows a predictable pattern for marking the future tense in Spanish but is incorrect, then mediation guides them to remembering that dormir is an irregular verb. The mediation sequence continues until either the learner selects the correct response or the final prompt is given, which reveals the answer and explains why it is correct. In this way, the C-DA procedure follows the implicit-to-explicit sequence of mediation and embeds a degree of flexibility in responding to the nature of a learner’s error that is inspired by human-mediated DA. 

To date, Randall and Urbanski’s (2023) C-DA targets a set number of features of Spanish grammar and includes either multiple-choice items, such as the one described above, or ‘sentence-builder’ items in which students must combine elements to construct a grammatical sentence. For example, to produce the sentence Do we take music lessons? Randall and Urbanski explain that each word represents a slot to be filled, and doing so correctly requires knowledge of syntax (e.g., placing subject pronouns and direct objects in an interrogative construction) as well as appropriate verbal inflection (e.g., take rather than takes, and do rather than does). Again, each mistake a learner makes opens a mediational sequence to help them identify the nature of their error and attempt to correct it. As the authors observe, a wider variety of language features can certainly be added to the C-DA program, and as technology continues to advance, greater flexibility in mediation will likely be possible. 

Of course, C-DA offers only one path for integrating mediation, a particular quality of teaching, into assessment procedures for the purpose of identifying where subsequent instruction should be focused to optimally guide learner development. Other avenues, including ones mentioned here (e.g., DA with groups of learners, its use in both formal assessment contexts and classrooms), should also be pursued to fully realize the implications of Vygotskian theory for education. It is in this way that, rather than a bridge too far, we might understand that the gap between assessment and teaching need not exist at all. 

References 

Antón, M. (2009). Dynamic assessment of advanced second language learners. Foreign Language Annals, 42(3), 576–598. 

Haywood, H. C., & Lidz, C. S. (2006). Dynamic assessment in practice: Clinical and educational applications. Cambridge University Press. 

Levi, T. (2015). Towards a framework for assessing foreign language oral proficiency in a large-scale test setting: Learning from DA mediation examinee verbalizations. Language and Sociocultural Theory, 2(1), 1–24. 

Poehner, M. E. (2008). Dynamic Assessment: A Vygotskian approach to understanding and promoting L2 development. Springer. 

Poehner, M. E., Infante, P., & Takamiya, Y. (2018). Mediational processes in support of learner L2 writing development: Individual, peer, and group contexts. Journal of Cognitive Education and Psychology, 17(1), 112–132. 

Poehner, M. E. & Wang, Z. (2021). Dynamic assessment and second language development. Language Teaching, 54, 472–490. 

Randall, T. S. & Urbanski, K. (2023). Development of a computerized dynamic assessment program for second language grammar instruction and assessment. Language and Sociocultural Theory, 10(1), 50–81. 

Vygotsky, L. S. (2012). Thought and Language. MIT press.