Nonintentional analogical inference in text comprehension

We present findings suggesting that analogical inference processes can play a role in fluent comprehension and interpretation. Participants were found to use information from a prior relationally similar example in understanding the content of a later example, but they reported that they were not aware of having done so. These inference processes were sensitive to structural mappings between the two instances, ruling out explanations based solely on more general kinds of activation, such as priming. Reading speed measures were consistent with the possibility that these inferences had taken place during encoding of the target rather than during the later recognition test. These findings suggest that analogical mapping, though often viewed as an explicit deliberative process, can sometimes operate without intent or even awareness.     

Altering comprehension: The effect of biasing titles on text comprehension

Titles can alter the comprehension of a text by affecting the selection of information from a text and the organization of this information in memory. Text comprehension is assumed to involve an organizational process that results in the formation of a text base, an ordered list of semantic units—propositions. The text base can be used as a retrieval scheme to reconstruct the text. Procedures for assigning propositions as more relevant to some themes as compared to other themes are developed and applied to texts. Texts with biasing titles were used in an experiment to demonstrate that immediate free recall is biased toward the theme emphasized in the title. The comprehension process which is guided by the text’s thematical information is described.     

Phonemic effects in reading comprehension and text memory

Two experiments investigated the role of phonemic information in adult reading comprehension and replicated the visual tongue-twister effect in a new paradigm–a modified probe memory task. College students took longer to read sentences that repeated word initial consonants (tongue-twisters) than matched control sentences. Equally important, subjects also took longer to respond to probe words from tongue-twisters. Slower response times in both the sentence reading task and the probe memory task indicate that the tongue-twister effect is indeed phonemic in nature and that phonemic information is used in memory during comprehension.     

Robust inference processes in expository text comprehension

Expository text offers particular challenges to the reader because of the abstract and unfamiliar concepts that it presents and its distinctive structure. The present study had three interrelated aims: (1) It examined the impact of appropriate connectives on the reader's derivation of causal bridging inferences from expository text. (2) It scrutinised texts longer than the ones that we had previously examined (Singer, Harkness, & Stewart, 1997), which in turn made it possible to (3) evaluate the impact of position in the text on inference processing. In Experiments 1a and 1b, a joint profile of target reading times and inference answer times indicated that the inspected inferences reliably accompanied reading only in the presence of appropriate causal connectives. The connective-present conditions of Experiment 1 replicated our previous findings using shorter texts (Singer et al., 1997). Experiment 2 indicated that these inference processes are unaffected by text position. We interpreted these findings with reference to the inference validation model (Singer, Halldorson, Lear, & Andrusiak, 1992).     

A computational model for simulating text comprehension

In the present article, we outline the architecture of a computer program for simulating the process by which humans comprehend texts. The program is based on psycholinguistic theories about human memory and text comprehension processes, such as the construction-integration model (Kintsch, 1998), the latent semantic analysis theory of knowledge representation (Landauer & Dumais, 1997), and the predication algorithms (Kintsch, 2001; Lemaire & Bianco, 2003), and it is intended to help psycholinguists investigate the way humans comprehend texts.     

Lexical processing during saccades in text comprehension

We asked whether people process words during saccades when reading sentences. Irwin (1998) demonstrated that such processing occurs when words are presented in isolation. In our experiment, participants read part of a sentence ending in a high- or low-frequency target word and then made a long (40°) or short (10°) saccade to the rest of the sentence. We found a frequency effect on the target word and the first word after the saccade, but the effect was greater for short than for long saccades. Readers therefore performed more lexical processing during long saccades than during short ones. Hence, lexical processing takes place during saccades in text comprehension.     

Resolution of activated background information in text comprehension

The present experiments explored the resolution of activated background information in text comprehension. In Experiment 1, participants read passages that contained an elaboration section that was either consistent or qualified (inconsistent but then corrected to be consistent) with respect to the subsequently presented target sentence (see O'Brien et al., 1998). However, the experiment used two target sentences, and several filler sentences were inserted between the first and second target sentence. The results showed that the reading times for the first target sentence in the qualified elaboration version were significantly longer than those in the consistent elaboration version. These were consistent with O'Brien's study, and further indicated that the basic process captured by the memory-based view appears to generalize to the Chinese reader better than does the here-and-now view. More importantly, the results showed that the reading times for the second target sentence in the qualified elaboration version were as long as those in the consistent elaboration version. These further indicated that the activation of background information not only maintained the coherence of the text, but also allowed for the relevant information to be updated, resulting in a unified information set. When the information was reactivated during ongoing reading, it would be in the form of unified information. In Experiment 2, the first target sentence in each passage from Experiment 1 was converted to a filler sentence, and the second target sentence became the target sentence. The results of Experiment 2 showed that the reading times for the target sentence in the qualified elaboration version were significantly longer than those for the consistent elaboration version. These indicated that the delay between the target sentences and the elaboration section was not responsible for the lack of differences in Experiment 1, and confirmed the conclusion of Experiment 1.     

The processing of cohesion devices in text comprehension

Summary The cognitive operations involved in the processing of surface-cohesion devices for the construction of a coherent mental representation is a major issue in text comprehension. An experiment was carried out with two narratives presented in two versions: a high-cohesion version and a low-cohesion version derived from the high version with the use of several devices — two anaphoric markers, changes in the temporal connective and word order, omission of the thematic sentence - without modification of the text content. The subjects read and immediately recalled a high-cohesion text and a low-cohesion text. The results showed that lowering cohesion produced an insignificant increase (8%) in reading time, but a highly significant decrease (25%) in recall performance. It appears that the subjects did not execute the processing required by the cohesion devices. The results are discussed with respect to models of sentence comprehension in comparison with text comprehension and metacognitive aspects of reading comprehension.     

Aging and the impact of causal connections on text comprehension and memory

Abstract

The hypothesis that aging is associated with an increased dependence on text-based organizational cues to support recall was investigated in two experiments. In the first, young and older adults read pairs of sentences that varied in their causal coherence. Recall for the second sentence in each pair was then tested using the first as a cue. the pattern of results suggested that young adults were likely to spontaneously infer causal connections between sentences when none was provided, whereas older adults were more likely to depend upon text-based connections to establish coherence and support memory. Experiment 2 supported this conclusion by demonstrating that the recall performance of older adults benefited more than did that of young and middle-aged adults when participants (a) were prompted to produce a link between sentences or (b) actually produced an integrative response that established a causal connection between the first and second sentences.     

Monitoring of comprehension: The role of text difficulty in metamemory for narrative and expository text

The effect of text difficulty on metamemory for narrative and expository text was investigated. In Experiment 1, we found an interaction between type of text and type of question (thematic or detailed). For readers of narrative texts, correlations between predicted and actual performance were highest for detailed questions, but this pattern was reversed for readers of expository texts. Next, text difficulty was explored as a possible factor affecting metamemory accuracy. In Experiments 2 and 3, metamemory accuracy was a nonmonotonic function of text difficulty. Subjects made remarkably accurate predictions of future performance (meanG > .6) for both narrative and expository texts that were of intermediate difficulty (approximately a 12th-grade reading level). We propose anoptimum effort hypothesis, predicting greatest metamemory accuracy when the texts are of intermediate difficulty.     

Coactivation and comprehension: Contribution of text variables to the illusion of knowing

Memory &; Cognition - When a reader’s self-assessment of comprehension is high, but an objective assessment reveals that comprehension is faulty, an illusion of knowing has occurred. This...     

A program for comprehension monitoring of text using HyperCard for the Macintosh

We describe a program using HyperCard for the Macintosh that allows the collection of several different types of text processing data in a realistic, natural, and unobtrusive experimental setting. The program allows a sentence-by-sentence presentation of text and allows unconstrained movement forward and backward through the text. Subjects can, at any point during reading, look back at individual sentences presented previously, or review all previously presented sentences at once. All sentence readings and rereadings are timed, and the sequence of readings, rereadings, and reviews is recorded. Following passage reading, questions are asked to examine comprehension and memory of the passage. The program is well suited for the examination of comprehension evaluation and regulation in readers, as well as for the examination of other comprehension and memory processes.     

On the relation between representations constructed from text comprehension and transitive inference production

Deductive inference production from texts is a process considered to involve either the construction of an integrated mental model or the step-by-step coordination of propositional representations of the sentences. These alternative hypotheses were tested in 3 experiments using a set inclusion task paradigm in which participants had to recall the premises and to evaluate transitive inferences. Contrary to what is known about linear ordering relations, order of recalls and reaction times provide evidence that the encoding of set inclusion relations does not result in an integrated representation. These results suggest that the mental models theory needs to take account of the nature of the relation to be represented if it is to become a general theory of reasoning.     

Effects of text difficulty and adults' age on relative calibration of comprehension

Weaver and Bryant (1995) proposed the optimum effort hypothesis, suggesting that undergraduate students were better able to predict comprehension when text materials matched their reading level (grade 12) as opposed to being too easy or too difficult (Weaver and Bryant did not assess the actual reading level of their participants). In the study, we examined the reading level and accuracy of performance prediction of both younger and older adults using Weaver and Bryant's materials. Regardless of our participants' high reading levels (grade 14 and above), they still predicted performance best when texts were written at around the grade 12 level, failing to support the optimum effort hypothesis.     

Effects of an intervention in active strategies for text comprehension and recall

The aim of this study was to investigate the effects of an intervention program to promote active text-processing strategies (main-idea identification and summarization) at two developmental levels (12- and 16-year-olds). The independent variables were training condition (experimental and control) and school level (7th and 10th grades). Several measures were taken as dependent variables: reading span, reading time, construction of macrostructure, and structural recall. The hypothesis claimed that training would increase comprehension and recall significantly. Furthermore, as a result of the training program, a reduction in developmental differences in the experimental groups at posttest was also expected. Results supported the predictions, showing a significant improvement in the experimental groups' reading comprehension and recall. These results are discussed in terms of the importance of active and self-controlled strategies for text comprehension and recall.