Academic journal article Educational Technology & Society

A Genre-Based Perspective for the Development of Communicative Computerized Adaptive Tests

Academic journal article Educational Technology & Society

A Genre-Based Perspective for the Development of Communicative Computerized Adaptive Tests

Article excerpt


This paper discusses the communicative potentials of Computerized Adaptive Testing. The study is based on a model that offers a set of independent communicative concepts to describe the Genre of an interactive application. This model will be the starting point to analyze the stages of the interaction cycle that are typically inherent to every generator of computerized adaptive tests. In each of these stages, the authors present the most appropriate communication structures that should support the intentions of the developers of this kind of software. The paper finishes giving a synthesis of the interactive potential of these communication stages from a Genre perspective, and discussing the compatibility between the proposed approach and some model-based methods that are currently used during the design of interactive applications.


Genre, Communication structures, Interactive applications, Software development, Computerized adaptive testing


The systems that use Computerized Adaptive Tests (CATs) aim to measure the maturity or knowledge level of a student in a specific domain by emulating the intelligent behavior of human evaluators. In other words, these tests dynamically apply the questions that are more informative for that purpose (Wainer, 2000). The authors have some experience in this area within the language learning domain, particularly in the Basque language field (Lopez-Cuadrado, Armendariz, & Perez, 2005; Lopez-Cuadrado, Perez, Arruabarrena, Vadillo, & Gutierrez, 2002).

The issues concerning the administration of items at run-time, which depends on previous responses, are well documented in literature. For instance, topics based on experimental protocols and psychometric measures such as stabilization of domain criteria, identification of existing dependences among them, classification of items before they feed the item bank, test validity, precision and security, item selection rules, and scoring methods are first and foremost treated in research and development projects (Olea & Ponsoda, 2003; van der Linden & Glas, 2000; Wainer, 2000). However, there is little work centered on the communicative dimension of CATs, in other words, how the successive stages of the assessment activity communicate with the receiver, that is, the examinee.

This paper works on a new a research line to propose guidelines to make easier the explicit description of the communicative dimension. This component, which belongs to the human-computer interaction, distinguishes between vertical interaction, i.e. studies of the dynamic offered to the final user, and horizontal interaction, i.e. studies about the presentation and also about the manipulation of the user-interface (Nogier, 2005). The former is derived from task-based models, such as ConcurTaskTrees (Paterno, 1999) or Hierarchical Task Analysis (Annett, 2003), which provide a scenario specification for interactions; and the latter is derived from graphic charts and ergonomic patterns (Lynch & Horton, 2005). The resulting interactive system is generally stabilized by means of participative and iterative approaches combined with prototyping techniques.

In the particular case of CAT generators, the vertical interaction is accepted and established by consensus (Thissen & Mislevy, 2000), as will be shown in section 3 by an algorithmic representation of their behavior. Actually, the horizontal interaction is the one that gathers less results and specific recommendations. Precisely, the contribution of this paper lies in this field.

While specifications of a horizontal interaction are basically centered on graphic charts and ergonomic patterns, this paper proposes the Genre model as perspective for complementary studies. Genre theory states explicitly that the development of a software product within a specific Genre--for instance, educational--makes the author to use communication structures--such as an exercise or course--that the receiver cognitively associates to the idea he or she has about the specific Genre (Latapy, Lopisteguy, & Dagorret, 2004). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.