Linguistic, textual and individual factors affecting children's comprehension of anaphors and connectives in reading

Linguistic, textual and individual factors affecting children's comprehension of anaphors and connectives in reading

First Author: Barbara Arfe -- University of Padova
Additional authors/chairs: 
Jane Oakhill, Univesity of Sussex
Keywords: Comprehension difficulty, Reader characteristics, Reading comprehension, anaphor, Knowledge of Connectives
Abstract / Summary: 

Readers use anaphors and connectives to integrate the meaning of sentences in a coherent mental representation of the text. However, the linguistic, individual and textual factors affecting the understanding of local coherence relations are still little understood, especially in young readers. This symposium brings together studies that were conducted in different languages, and in different populations (typically developing readers and poor readers with and without hearing loss) to further this understanding. The first two contributions examine the development of pronoun resolution and the processing of anaphors during reading in typically developing readers and in poor readers with intellectual disability and deafness. The third contribution provides an overview of the developmental understanding of different coherence relations. The final two contributions examine the textual, linguistic and individual factors that contribute to, and/or constrain the understanding of temporal (and causal) connectives in typically developing and in poor readers with or without hearing loss.

Symposium Papers: 

The development of Taiwanese children’s pronoun resolution abilities during reading

First Author/Chair:Yuhtsuen Tzeng -- National Chung Cheng University
Additional authors/chairs: 
Chiu-Hua Huang, To Ko University

Purpose
The purpose of these studies was to examine the developmental patterns of Taiwanese children’s abilities to resolve pronoun during reading and how they were modulated by relevant factors.

Method
Short texts consisted of two sentences were constructed following such a general format: The first sentence always mentioned two characters and the second sentence referred to one of the characters by using a proper pronoun. Features of characters and pronouns varied by specific research aims. Comprehension of each text was tested immediately followed by a separate interrogative question. Primary school students read in a self-paced manner at normal speed. Sentence reading time or eye movement data were recorded.

Results
Older students were more accurate in answering the comprehension questions and read the second sentence faster. Their accuracy rate went above 90% for students beyond 4th grade when the pronoun carried gender cues. Students beyond 3rd grade demonstrated a first-mention preference if the pronoun is “他/ta/he”. This first-mention effect did not appear until 5th grade if the second sentence has zero anaphor. There is a recency effect for 2nd and 3rd graders if the pronoun is “她/ta/she”. A typical repetition penalty effect was observed for as young as grade 3. Pronoun resolution also modulated by the semantic relationships between sentences. High causal relations between sentences speed up reading time for zero pronoun sentences even for 2nd graders.

Conclusions
Many factors contribute to the success of children’s pronoun resolution and each factor affect different aspects at different developmental level.

Anaphor processing in poor readers with Intellectual Disability and Pre-locutive deafness

First Author/Chair:Inmaculada Fajardo -- Department of Developmental and Educational Psychology, University of Valencia (Spain)
Additional authors/chairs: 
V. Ávila, University of Valencia ; L. Gil, University of Valencia ; A. Ferrer, University of Valencia

Purpose
Previous studies suggest that in adult regular readers, repeated name anaphors would facilitate lexical access due to a repetition priming effect, but would also interfere with information integration (Huang, Hopfinger & Gordon, 2014; Ledoux et al., 2007), while in children and readers with difficulties, repeated names would benefit both processes in comparison with less explicit anaphor forms (pronoun and null anaphors) (e.g. Ehrlich & Remond, 1997; Rodríguez et al., 2010). We tested the effect of anaphor form (repeated, pronoun and null anaphors) during a reading comprehension task in Spanish for two different groups of poor readers (readers with intellectual disability (ID) and pre-locutive deafness (PD)) and a group of reading-level-matched readers with typical development (TD) (3rd grade reading level).

Method
The three groups of participants (nID=11; nPD=13 and nTD=16) were asked to read two-sentences texts with anaphors and to answer comprehension questions while their eye-movements were registered.

Results
Participants of the three groups performed fewer and shorter fixations during the first pass (early measures reflexing lexical access) of sentences containing repeated names and null anaphors than those containing pronoun anaphors. In addition, readers with ID made more regressions (later measure reflexing integration) to antecedent sentences when texts contained null versus pronoun anaphors.

Conclusions
These results, which suggest that more explicit anaphor forms would benefit poor readers, or at least not hinder, during both lexical access and integration, are discussed in relation to the design of simplified reading materials.

Establishing coherence using coherence markers: a developmental overview

First Author/Chair:Ted J.M. Sanders Sanders -- Utrecht University
Additional authors/chairs: 
Jacqueline Evers-Vermeul, Utrecht University

Purpose
Coherence markers are important cues for readers to see how sentences are related. We present a developmental overview: are connectives (because, but), cue phrases (as a result), and signalling phrases (the reason for this is…) beneficial to readers of all ages, and which role do they play in text processing?

Method
We review evidence from experimental studies on the role of coherence markers in reading comprehension, addressing off-line and on-line results from children, adolescents, and adults.

Results
Children show a gradual increase in their ability to benefit from connectives in texts (Cain & Nash 2011; Irwin and Pulver 1984; Yuill & Oakhill 1991). Even children in grade 4 benefit from explicit instruction on how to combine sentences with coherence markers (Williams et al. 2014). For adolescents, the addition of connectives leads to better answers to bridging inference questions about History, Economy and Biology texts, and narratives (Land et al. 2007; van Silfhout et al. 2014, 2015). For adults the presence of coherence markers results in better and faster recall (Lorch & Lorch 1986; Millis & Just 1994), faster response on verification tasks (Sanders & Noordman 2000), and higher scores on comprehension tasks (Degand & Sanders 2002; Kamalski et al. 2008; McNamara et al. 1996).

Processing results indicate that coherence markers serve as processing instructions: by providing information on the type of relation to be established, thereby restricting the number of inferences that can be made (Millis et al. 1993; Koornneef & Sanders 2013), they allow the reader to process upcoming information more quickly (Millis & Just 1994; van Silfhout et al. 2014, 2015).

Conclusions
We will address implications for education

The role of connectives in elementary school children’s comprehension of temporal relations between events during reading.

First Author/Chair:van den Broek Paul -- Brain and Education Lab, Leiden University
Additional authors/chairs: 
Josefine Karlsson, Brain and Education Lab, Leiden University; Linda Van Leijenhorst, Brain and Education Lab, Leiden University

Purpose
In two experiments we examined the variables that influence children’s comprehension of temporal relations between events in sentences with temporal connectives. We aimed to explain children’s comprehension difficulties from the interaction of text characteristics and children’s limited cognitive processing resources. To this end we examined effects of sentence chronology, position of the connective, position of the main clause, and recency of mention of the correct answer, as well as of individual differences in children’s executive functions.

Method
In experiment 1, 9-12 year old children (N=80) read compound sentences such as ‘Before Bart ate a cookie, he drank milk’ with the temporal connectives before or after, in a computerized task. For each pair they indicated which event happened first. In experiment 2, children (N-52) read the same sentences but indicated which event happened last. In both experiments we measured working memory and cognitive flexibility. Accuracy was analysed using repeated-measures ANOVAs.

Results
Across experiments, position of the main clause and recency of mention of the correct answer influenced children’s comprehension of temporal relations, whereas sentence chronology and connective position did not. Specifically, when main-clause position or recency of mention did not support correct comprehension, children made relatively many errors. Working-memory and cognitive-flexibility scores revealed that children with better executive functions made fewer errors.

Conclusion
These findings indicate that in 9-12-year-old children’s difficulties comprehending temporal relations between events arise when their limited cognitive-processing resources are taxed. Because these children are reading to learn, such comprehension difficulties may be especially detrimental to them.

Comprehension of temporal and causal connectives in poor comprehenders with and without hearing loss.

First Author/Chair:Arfé Barbara -- DPSS-University of Padova
Additional authors/chairs: 
Jane Oakhill, University of Sussex; Paola Benincasa, University of Modena and Reggio Emilia; Paul van den Broek, Leiden University; Elisabetta Genovese, University of Modena and Reggio Emilia

Purpose
Hearing poor comprehenders (PCs) and deaf or hard of hearing poor comprehenders (DHH/PCs) have problems with the understanding of causal and temporal connectives in texts. However, their performance in this task has been never compared before. We compared DHH/PCs and PCs’ understanding of temporal and causal coherence relations expressed by connectives (e.g. before, because or and) and examined the role plaid by linguistic knowledge (receptive vocabulary) and kind of coherence relation (causal or temporal).

Method
In Study 1, DHH/PCs (n=27, 7-11 years), age matched hearing PCs (n=31) and good comprehenders (GCs, n=32) had to choose among three sentences the causal or temporal sentence that best matched a depicted scene. The sentences differed only in their connectives (before, after, while, because, or so). In Study 2, the three groups were administered a revised version of the task, in which children had to choose between explicit causal or temporal connectives and the connective "and". Mixed ANOVAs compared the performance of the groups in the two experimental conditions (causal versus temporal relations). Vocabulary scores were used as covariate.

Results
PCs performed worse than GCs, but had a better understanding of explicit causal and temporal connectives than DHH/PCs. All groups showed better understanding of causal than temporal relations. The difference in connective understanding between DHH/PCs and PCs was not explained by their vocabulary knowledge.

Conclusions
The results suggest that DHH children’s difficulties in understanding connectives are not entirely explained by their poor linguistic (vocabulary) knowledge and depend on the coherence relation expressed.