COMPREHENSION ABILITY IN ADULT READERS

A.    INTRODUCTION

            Readers construct detailed mental representation of texts with relative ease. Reading is a complex, intellectual skill, requiring the coordination of multiple component processes. Three levels of the processes when readers are trying to make sense of the ideas: word level (to encode the printed word, sound-based representation, retrieve its meaning from memory), sentence level (involved in understanding “who did what to whom” in a sentence), and discourse level (reflects features of the real or imaginary world that the text describes).

            Poor comprehenders have difficulty making inferences to integrate ideas in a text, to answer question, and to identify main ideas and themes. Reading comprehension involves language specific processes as well as domain-general cognitive abilities-sensation, perception, attention, memory, and reasoning.

B.     INDIVIDUAL DIFFERENCES IN WORLD-LEVEL ABILITIES

            In beginning readers, word-identification skill is strongly related to measures of phonological awareness-explicit knowledge about the phonological structure of the language. The three task that are most commonly used to assess adult reader’s word-level abilities are: naming, phonological decision, and orthographic decision. Verbal efficiency theory suggests two ways that deficits in word-identification skill can influence comprehension performance: (1) Comprehension processes that depend on high quality, lexical representations, such as syntactic analysis. (2) Slow retrieval of lexical codes can compromise higher-level interpretive processes by consuming WM resource that would otherwise be devoted to these processes.

            Verbal efficiency theory emphasizes three characteristics: (1) word-level ability, (2) WM capacity, (3) print exposure.

C.    WHAT ARE THE STATUS OF COMPREHENSION PROBLEMS FOR POOR QUALITY REPRESENTATION?

            If poor comprehenders can quickly and accurately select context-appropriate senses of ambiguous words, their failure to make topic-related inference should be attributed factors other than poor quality sentence representations.           Both good and poor comprehenders recognized targets faster when they were preceded by primes from the same that different proposition. Only good comprehenders showed topic-priming effects-difficulty rejecting topic words that were related to the passages. Good and poor comprehenders showed differences in (a) their ability to integrate information from different parts of the story (b) their ability to elaborate their representation with topic-related information.

            In short, poor comprehenders have word-level and sentence-level processes that are accurate enough (a) to encode structural relations among concept in sentence (b) to use content in selecting the appropriate sense of an ambiguous word (c) to support reactivation of prior text information. Poor comprehenders appear to construct quality sentence representation.

D.    DO COMPREHENSION PROBLEMS RELATE TO SLOW WORD-LEVEL PROCESSING?

            A primary claim of verbal efficiency theory is that slow word-level processes can consume resources that would otherwise be devoted to higher-level interpretive ones. The researchers found that (1) word-decoding speed was a unique predictor of sentence processing, (2) working-memory capacity was not a reliable predictor of any level-1 coefficient. The coefficient associated with the number of function words in a sentence was predicted by WM capacity, but only when the other individual-differences factors were eliminated.

            In brief, verbal efficiency theory claims that slow and inaccurate word-level ability is associated with reading comprehension in adult readers as it is in children. Adult readers have word-level processes that are accurate enough for the construction of reasonably good sentence-level representation. Slow word-level processing appears to be predictive of reading comprehension, independent of other factors such as general verbal ability and WM capacity.

E.     INDIVIDUAL DIFFERENCES IN WORKING MEMORY

            WM is the theoretical construct used to refer to the system that is responsible for maintaining such information. To assess WM capacity, researchers used two measures: (a) the reading-span task [two activities: read aloud a set of unrelated sentences, presented one at a time, and recall the final work of each sentence once the entire set has been presented], (b) the operation-span task [participants perform simple arithmetic problems rather that read aloud].

1.      Limitations Due to Capacity Constraints

a.      The Capacity Theory of Comprehension

The capacity theory attributes individual differences in reading comprehension to variation in capacity, the total amount of activation available to the system. Studies that examine the relation between WM and discourse-level processes have found similar results: (a) high-span readers compared to low-span, are more accurate in finding the antecedent of a pronoun when the pronoun and its antecedent are separated by intervening sentences, (b) high-span readers are also more likely than low-span readers to show faster recognition of sentences that are thematically related.

b.      Separate-Sentence-Interpretation-Resource (SSIR) Theory

The foundation of The SSIR theory lies in neuropsychological data which concern the ability of patients to understand sentences containing complex syntactic structures. (1) One part of the WM system is specialized for analyzing syntactic structure used to determine sentence meaning [sentence interpretation]. (2) Another

part of the system is devoted to activities that involve conscious controlled processing [post-interpretive], these activities include: making inference to integrate ideas across sentences, using world knowledge in the interpretation of a text, remembering sentence content, and planning actions based on the meaning of sentences and texts. Like the capacity theory, the SSIR theory predicts a strong relation between WM capacity and discourse-level processing.

2.      Limitations Due to Poor Word-Level Ability and Insufficient Experience

a.      The Connectionist-Based Account

The connectionist-based account is based on connectionist approaches to language

processing. In the connectionist-based account, individual differences in performances on WM task arise from variation in two factors, (1) Individuals can vary with respect to basic sensory/perceptual abilities, primarily the ability to represent phonological information accurately. (2) Individuals can vary in reading experience. According to the connectionist-based account, variation in performance on span tasks is due to the same factors that influence all language tasks: word-level ability and print exposure.

b.      The Long-Term Working Memory (LTWM) Model

LTWM is a mechanism based on skilled storage and retrieval in long-term memory. Individuals who are skilled in mental calculated have large WM capacities as measure by digit span but do not show large capacities for other types of materials. The LTWM model is different from the connectionist-based account in that it includes a traditional view of a fixed-capacity WM system.

F.     INDIVIDUAL DIFFERENCES IN SUPPRESSION ABILITY

            Suppression ability (cognitive inhibition) is an individual’s skill at ignoring or inhibiting distracting information and overcoming interference from a powerful response. Suppression diminished activation of the trace when their content is unrelated to the structure. Suppression is an automatic inhibitory mechanism or a controlled strategic one. In the controlled-attention view, variation in performance on both comprehension tasks and complex span tasks are due to individual differences in the ability to control attention, including the ability to suppress irrelevant information.           

G.    INDIVIDUAL DIFFERENCES IN PRINT EXPOSURE

            Comprehension skill is likely influenced by print exposure in at least three ways: (1) Individuals who read often are more likely to learn about rare words than are individuals who read seldom [vocabulary growth is likely to be accelerated in individuals who read often], (2) individuals are more likely to encounter complex syntactic structures in print than in speech, and (3) individuals who read often are likely to acquire ore world-knowledge than individuals who read seldom. Print exposure gives a strong effect on background knowledge via its effect on discourse-level processing.

H.    INDIVIDUAL DIFFERENCES IN BACKGROUND KNOWLEDGE

            High-knowledge readers construct discourse models [recall, problem-solving, generalization, and knowledge-based inferences] in which text ideas are integrated with each other and with a large network of relevant prior knowledge. Low-knowledge readers construct text representations that are coherent at the sentence-level, but they lack the knowledge necessary to construct coherent discourse models. Low-knowledge readers can recognize ideas from a text, but can not use their representation to perform tasks that require conscious, reflective access to a discourse model.

            Background knowledge is essential for building retrieval structures in LTWM that expand a reader’s ability to hold large amounts of information in an accessible form. Background knowledge is facilitated by print exposure. Individuals who read often are exposed to more information about the world and they are more likely to create coherent discourse models that expand their knowledge base.

original resource: Suparman, Ujang. 2010. Psycholinguistics: The Theory of Language Acquisition. Bandung: Arfino Raya.