Complexidade textual de itens de avaliação em larga escala: proposta de um mecanismo analítico
The growing concern with the improvement of educational quality has focused on the process of large-scale evaluations in Brazil, allowing a reformulation through the Reference Matrix that includes the competences and the measured skills. Against this background, this thesis has the primary object...
Na minha lista:
| Autor principal: | |
|---|---|
| Outros Autores: | |
| Formato: | doctoralThesis |
| Idioma: | pt_BR |
| Publicado em: |
Brasil
|
| Assuntos: | |
| Endereço do item: | https://repositorio.ufrn.br/jspui/handle/123456789/28075 |
| Tags: |
Adicionar Tag
Sem tags, seja o primeiro a adicionar uma tag!
|
| Resumo: | The growing concern with the improvement of educational quality has focused on the process
of large-scale evaluations in Brazil, allowing a reformulation through the Reference Matrix that
includes the competences and the measured skills. Against this background, this thesis has the
primary objective of proposing an analytical mechanism to identify, based on criteria related to
the structure of the text, the degree of Textual Complexity (TC) of a large-scale evaluation item.
Specifically, it is aimed to: (a) discuss the notion of textual complexity; (b) to propose criteria
related to the compositional dimension of the text to measure the degree of complexity of a
large-scale evaluation items; (c) apply the notion of cognitive domain to the analysis of the
complexity of evaluation items. It was adopted an objective structural conception of TC that
concerns to the relation of compositional attributes of a text that make it more or less complex.
It is a different perspective from the ones that base the studies that have been done about TC,
which usually take as a base the lexicon and the syntactic structure. In the project undertaken
here, we take into account the basic item procedure representative of large-scale evaluation
items, whose sample therefore serves to demonstrate and discuss the possibility of applying the
proposed analytical model. On this basis, we start from the hypothesis that TC can be measured
by criteria related to the compositional and structural dimension of the text. To fulfill this goal,
we started from two approaches of textual studies to establish categories of TC analysis.
Sociodiscursive Interactionism (SDI) pointed to the types of discourse as a category. The
Interactive Textual Perspective (ITP) presented the discursive topic. In addition to these
approaches, we consider Bloom's Taxonomy (BT), which indicated properties related to the
cognitive domain mobilized by the text. As the research is centered on the understanding and
explanation of this TC, it is a qualitative work in approach and applied in nature. Considering
also that we intend to discuss TC in order to make it more explicit and build hypotheses, we are
in the field of exploratory research. The proposed analytical mechanism then comprises two
criteria: the compositional dimension (discourse types and discursive topic) and the cognitive
dimension (cognitive domain). Based on the types of discourse, the following properties are
foreseen: interactive discourse, theoretical discourse, interactive report, narration. Regarding
the discursive topic, the properties are: one or more of a topic, topic with or without unfolding,
implicit or explicit relation of concern. Finally, the properties of the cognitive domain are:
knowledge, comprehension, application, analysis, synthesis and evaluation. Based on these
properties, complexity scales are identified according to which an item can be considered, from
the textual point of view, more or less complex. The feasibility of this mechanism was applied
to a corpus composed of 24 items of the ENEM 2016 test, from the answers of more than
300,000 students in the application Quero Minha Nota! (I Want My Grade!), which estimates
the exam grade, TUNEDUC, a startup, specialist in high school educational data, namely:
Languages, Codes and their Technologies; Mathematics and its Technologies; Humanities and
their Technologies; Natural Sciences and their Technologies, which listed the 12 easiest items
and the 12 most difficult items. Thus, it was concluded that textual complexity is an issue that
should be considered in the broad discussion about how to elaborate and apply a large-scale
evaluation item. The results of the research indicated that, based on the dimensions and criteria
used, the items do not confirm this classification of easy and difficult, as the INEP claims by
TRI, but rather an item is low and of high complexity. |
|---|