Assessment

The term dynamic assessment (DA) refers to an assessment, by an active teaching process, of a child's perception, learning, thinking, and problem solving. The process is aimed at modifying an individual's cognitive functioning and observing subsequent changes in learning and problem-solving patterns within the testing situation. The goals of the DA are to: (a) assess the capacity of the child to grasp the principle underlying an initial problem and to solve it, (b) assess the nature and amount of investment (teaching) that is required to teach a child a given rule or principle, and (c) identify the specific deficient cognitive functions (i.e., systematic exploratory behavior) and non-intellective factors (i.e., need for mastery) that are responsible for failure in performance and how modifiable they are as a result of teaching. In contrast, the term static test (ST) generally refers to a standardized testing procedure in which an examiner presents items to an examinee without any attempt to intervene to change, guide, or improve the child's performance. A static test usually has graduated levels of difficulty, with the tester merely recording and scoring the responses.

DA is usually administered to children who demonstrate some learning disability, low scores on standardized tests, or some emotional or personality disturbance. Very frequently it is given to children coming from a low socioeconomic or culturally different background. The differences between the ST and DA approaches derive from different philosophical perspectives: ST is related to passive acceptance (acceptance of a child's disability and accommodation of the environment to fit these disabilities), while DA is based on active modification (active efforts to modify the child's disabilities by intensive mediation and the establishment of relatively high cognitive goals).

DA development has been motivated by the inadequacy of standardized tests. The inadequacy can be summarized in the following points: (1) Static tests do not provide crucial information about learning processes, deficient cognitive functions that are responsible for learning difficulties, and mediational strategies that facilitate learning. (2) The manifested low performance level of many children, as revealed in ST, very frequently falls short of revealing their learning potential, especially of those identified as coming from disadvantaged social backgrounds, or as having some sort of learning difficulty. Many children fail in static tests because of lack of opportunities for learning experiences, cultural differences, specific learning difficulties, or traumatic life experiences. (3) In many static tests children are described in general terms, mostly in relation to their relative position of their peer group, but they do not provide clear descriptions of the processes involved in learning and recommendations for prescriptive teaching and remedial learning strategies. (4) Static tests do not relate to non-intellective factors that can influence individuals' cognitive performance, sometimes more than the "pure" cognitive factors. Nonintellective factors (i.e., intrinsic motivation, need for mastery, locus of control, anxiety, frustration, tolerance, self-confidence, and accessibility to mediation) are no less important in determining children's intellectual achievements than are the "pure" cognitive factors. This is especially true with individuals whose emotional or motivational problems interfere with their cognitive performance.

In comparison with ST, DA is designed to provide accurate information about: (a) an individual's current learning ability and learning processes; (b) specific cognitive factors (i.e., impulsivity, planning behavior) responsible for problem-solving ability and academic success or failure; (c) efficient teaching strategies for the child being studied; and (d) motivational, emotional, and personality factors that affect cognitive processes.

Lev Vygotsky's concept of a zone of proximal development (ZPD) and Reuben Feuerstein's theory of mediated learning experience (MLE) served as the main conceptual bases for most of the DA elaboration. The ZPD is defined as the difference between a child's "actual developmental level as determined by independent problem solving" and the higher level of "potential development as determined through problem solving under adult guidance or in collaboration with more capable peers" (Vygotsky, p. 86). MLE interactions are defined as a process in which parents or experienced adults interpose themselves between a set of stimuli and a child and modify the stimuli for the developing child. In a DA context, the examiner mediates the rules and strategies for solving specific problems on an individual basis, and assesses the level of internalization (i.e., deep understanding) of these rules and strategies as well as their transfer value to other problems of increased level of complexity, novelty, and abstraction.

The Nature of Dynamic Assessment

DA is meant to be a complement to standardized testing, not a substitute for it. It is presented as a broad approach, not as a particular test. Different criteria of change are used in DA: pre-to post-teaching gains, amount and type of teaching required, and the degree of transfer of learning. The choice to use change criteria to predict future cognitive performance (as well as predicted outcome of intervention programs) is based on the belief that measures of change are more closely related to teaching processes (by which the child is taught how to process information), than they are to conventional measures of intelligence. The major differences between DA and conventional tests in regard to goals, testing processes, types of instruments, test situations, and interpretation of results, are presented in Table 1.

Using DA. Clinical experience has shown that it is most useful to use DA when standardized tests yield low scores; when standardized tests hover around margins of adequacy in cognitive functioning; when there are serious discrepancies between a child's test scores and academic performance; when a child comes from a low socioeconomic or culturally or linguistically different background; or when a child shows some emotional disturbance, personality disorder, or learning disability.

Reliability of DA. One of the objectives of DA is to change an individual's cognitive functioning within the testing context so as to produce unreliability among test items (i.e., lack of consistency between repeated responses). DA reliability is usually assessed by interrater agreement (two or more observers rate the child's behavior) regarding the child's cognitive performance, mediation (teaching) strategies required to change the child's functioning, cognitive functions (i.e., level of impulsivity, planning behavior) that affect performance, and motivational-emotional factors. Such test reliability has been demonstrated with learning disabled and educable mentally retarded (EMR) children. Overall inter-rater agreement for the type of intervention (mediation) required to change a child's performance for deficient cognitive functions, such as impulsivity, lack of planning, and lack of systematic behavior, has been shown to be about 89 percent. For different cognitive tasks, different profiles of deficient cognitive functions have been observed and different types of teaching can be applied.

Current Research

Educational perspectives. Previous research has shown that standardized IQ tests underestimate the cognitive ability of children from low socioeconomic settings, from minority groups, and children having learning difficulties. Criteria of change (i.e., pre-to post-teaching gains on a test), as measured by DA, have been found to be more powerful in predicting academic performance, more accurate in prescribing individualized educational plans and specific cognitive interventions, and better able to distinguish between different clinical groups than ST scores. David Tzuriel and Pnina Klein, using the Children's Analogical Thinking Modifiability (CATM) test, showed that the highest pre-to post-teaching gains were found among children identified either as disadvantaged or advantaged, compared with children with special education needs and mentally retarded children. Higher levels of functioning, for all groups, were found on the CATM than on the Raven's Colored Progressive Matrices (RCPM)–when the latter was given as a standardized test–especially when comparing performance on analogy items of the RCPM versus problems on the CATM. The advantaged and disadvantaged children scored 69 percent and 64 percent, respectively, on the CATM, compared with 39 percent and 44 percent on the RCPM. The effects of teaching were more articulated in difficult tasks than in easy ones.

TABLE 1

Findings with the Children's Inferential Thinking Modifiability (CITM) and the Children's Seriational Thinking Modifiability (CSTM) tests indicate that children from minority groups or disadvantaged background have an initial lower level of functioning than children from mainstream groups or an advantaged background. After a teaching phase, however, they showed higher levels of gain and narrowed the gap. The gap between the two groups was also narrower in a transfer phase consisting of more difficult problems. The degree of improvement was higher in high-complexity problems than in lowcomplexity problems.

In several studies DA was found to verify the distinction between cultural deprivation and culturaldifference. Tzuriel, following Feuerstein, differentiated between those who function poorly as a result of cultural differences and those who have experienced cultural deprivation. The DA approach, in this respect, offers a solution not only for its differential diagnostic value, but also for its potential prescriptive remediation of deficiencies and its enhancement of learning processes.

For certain DA measures, significant positive correlations have been found between the level of difficulty of an item and the level of improvement on that item, and DA post-teaching scores have been shown to be better predictors of academic achievement than static scores. In addition, a higher prediction value was found among children with high learning potential than among children with average learning potential. Findings of many studies raise heavy doubts, especially with low functioning groups, about the ability of ST scores to represent accurately an individual's ability and to serve as indicators for future intervention and change.

Evaluation of Cognitive Education Programs

Dynamic assessment has also been used to evaluate cognitive education programs designed to develop learning and thinking skills. Given that one of the major goals of these programs is advancing learning to learn skills, it is essential that the change criteria in DA be assessed several studies have shown that experimental groups who received any one of a number of cognitive education programs (e.g., Bright Start, Instrumental Enrichment, Peer-Mediation with Young Children) attained higher pre-to post-teaching gains on DA tests than did control groups. The DA scores depicted the effects of the intervention better than ST scores did.

Developmental Perspective

Developmental research using DA has focused on predicting learning ability by assessing the quality of parent–child interactions, specifically mother–child mediated learning experience (MLE). MLE interactions are defined as an interactional process in which parents, or substitute adults, interpose themselves between a set of stimuli and the child and modify the stimuli for the developing child. The mediator modifies the stimuli by focusing the child on their characteristics, by arousing curiosity, vigilance, and perceptual acuity in the child, and by trying to improve and/or create in the child the cognitive functions required for temporal, spatial, and cause-effect relationships. Major findings have been that children's post-teaching scores are more accurately predicted by MLE mother–child interactions than by ST scores and that mediation for transcendence (expanding an experience by making a rule or principle, generalizing an event beyond the concrete experience) has emerged as the most powerful predictor of children's ability to change following teaching. These findings support the hypothesis that mother–child mediation strategies are internalized and used later in other learning contexts. Children whose mothers used a high level of mediation for transcendence internalized the mechanism and used it in other learning contexts where they needed this type of mediation. Findings of several studies confirm the hypothesis that MLE interactions, conceptualized as the proximal factor of cognitive development (i.e., directly explaining cognitive functioning), predicted children's cognitive change, whereas distal factors (i.e., SES, mothers' IQ, child's personality orientation, mother's emotional attitudes toward the child) did not predict cognitive change in children.

Future Research

In spite of the efficacy of DA, some problems exist. First, DA takes more time to administer and requires more skill, better training, more experience, and greater effort than ST. A cost-effectiveness issue is raised by psychologists, educators, and policymakers who are not convinced that the information derived from DA is worth the investment required to get it, and that the information acquired will then be used efficiently to enhance specific learning strategies and academic achievements.

Second, the extent to which cognitive modifiability is generalized across domains needs further investigation. This issue has practical implications for the designing of tests and mediational procedures. Third, validation of DA is much more complex than validation of ST because it has a broader scope of goals (assessing initial performance, deficient cognitive functions, type and amount of mediation, nonintellective factors, and certain parameters of change). In validating DA one needs to develop criteria variables that measure changes that are due to a cognitive intervention.

Finally, the literature is replete with evidence showing a strong relation between IQ (an ST measure) and school achievement (r = .71). This means that nearly 50 percent of the variance in learning outcomes for students can be explained by differences in psychometric IQ. However, three extremely important questions remain: (1) What causes the other 50 percent of achievement variance? (2) When IQ predicts low achievement, what is necessary to defeat that prediction? and (3) What factors influencing the unexplained variance can help to defeat the prediction in the explained variance?