Deepening Comprehension: Exploring the Effects of a Book-Reading Intervention on Children's Inferential Thinking

• In Event: Children's Conceptual Knowledge, Thinking, and Teacher Talk: A Fellowship of Unsung Heroes in Language Comprehension

Abstract

Theoretical Framework
Recent research has shown that inferential thinking is critical to reading comprehension beyond other oral language skills, word recognition, and IQ (Lepola, Lynch, Kiuru, Laakkonen, & Niemi, 2016; Lepola, Lynch, Laakkonen, Silvén, & Niemi, 2012; Oakhill & Cain, 2012). Preschoolers have these capacities, such as drawing inferences and identifying causal connections in narratives (Kendeou, van den Broek, White, & Lynch, 2009; van Kleeck, 2008). The quality of talk by teachers is related to children’s comprehension and thinking (Collins, 2016). Despite the potential of early contexts, such as storybook reading, for developing such capacities, adults engage in more literal than inferential talk (Price, van Kleeck, & Huberty, 2009). Inferential thinking has received little attention in intervention research (Dunst, Williams, Trivette, Simkus, & Hamby, 2012) and few studies have articulated content for teacher training on children’s inferential thinking.

Objectives
Objectives are to examine the efficacy of a storybook reading intervention with focused support for inferential thinking on children’s literal and inferential story comprehension. Additionally, we sought to identify patterns of change in children’s comprehension, especially inferential thinking, across the year-long intervention.

Data Sources
This project analyzes data from an experimental study designed to test the effectiveness of a storybook reading and play intervention on children’s vocabulary acquisition.

Methods
Participants (N=199) were preschool children (Mage = 4 years, 6 months; 52% female) and teachers from 10 Head Start and 10 state-funded PreK classrooms in two urban cities in the northeastern and southern United States, respectively.

Children participated in whole group read-alouds implemented by the classroom teacher who employed vocabulary support strategies taught as part of the project. To support comprehension, teachers used scripted guidance for literal and inferential questions to ask during each reading. As part of the intervention, children also participated in play sessions focused on the vocabulary words. The project included three books, each of which was read four times over a period of four weeks. Books were unfamiliar, age-appropriate fictional narratives. Within approximately one week after the fourth reading and play session for each book, children took a story comprehension test consisting of four literal and four inferential questions that differed from the questions posed by the teacher during the readings. Scores on the literal and inferential questions were analyzed.

Results
Mixed-effects regression was used to predict the number of literal and inferential questions answered correctly by children across time (e.g., Book 1, Book 2, Book 3). The model includes random effects for subjects nested within classrooms which were nested within sites. Findings indicate that literal (β = .09, p < .05) and inferential scores (β = .67, p < .001) increased significantly over the year. Moreover, increases in inferential questions were significantly greater than increases in literal questions (β = .29, p < .001; see Figure).

Significance
Findings illuminate the malleability of children’s thinking and the potential of teacher training that includes language guidance for modeling and eliciting inferential thinking. Moderating variables, implications for educators, and the potential for change in classroom read-alouds will be discussed.

Authors

• Molly F. Collins, Vanderbilt University

• Kaitlin Herbert, Vanderbilt University

• Emily Hopkins, University of Scranton

• Jessica Lawson-Adams, Vanderbilt University

• Marcia Shirilla, University of Delaware

• Rebecca Dore, University of Delaware

• Molly Scott, Temple University

• David K. Dickinson, Vanderbilt University

• Roberta Michnick Golinkoff, University of Delaware

• Kathy Hirsh-Pasek, Temple University