Why Would You Say Goodnight to the Moon? Response of Young Intellectually Gifted Children to Lower and Higher Order Questions During Storybook Reading

Abstract

Research into the effect of questions asked during storybook reading in preschool settings has rarely investigated questions that elicit higher order thinking. In the current study, Blank et al.’s Four Levels of Abstraction were used to code teacher questions and child responses from 177 individual storybook reading sessions with eight intellectually gifted 3- and 4-year-old children. The aim was to investigate the level of cognitive response gifted preschoolers made to lower and higher order questions during shared book reading. As expected, lower order questions, the most frequent form of teacher questions, elicited mainly lower order responses. Significant cognitive correspondence was also found for higher order questions, which elicited higher order child responses 88% of the time. This suggests that higher order questioning would be a valuable addition to preschool storybook reading, particularly to extend the thinking of young intellectually gifted children.

Keywords

early childhood educational interventions evidence-based practice gifted children questioning

Asking higher order questions is often cited as an appropriate intervention strategy to challenge and extend the thinking of young intellectually gifted children (Harrison, 2003; Maker, 1986; A. Robinson, Shore, & Enersen, 2007; Sutherland, 2012). However, little research has investigated how young gifted children respond when asked higher order questions (Walsh & Kemp, 2013), that is, whether higher order questions do in fact elicit higher levels of thinking. It is important to evaluate this strategy as previous research with teachers across all levels of education (from preschool to high school) demonstrated that teachers tend to rely on lower order questions and ask very few higher order questions (Dickinson & Keebler, 1989; Dickinson & Smith, 1994; Dillon, 1982; Hindman, Wasik, & Erhart, 2012; S. R. Mills, Rice, Berliner, & Rousseau, 1980; Siraj-Blatchford & Manni, 2008; Zucker, Justice, Piasta, & Kaderavek, 2010). If evidence can be found that asking higher order questions does indeed lead to higher level responses, especially in the context of shared book reading (SBR), implementing this change in teacher behavior could be highly beneficial for young gifted children.

Higher order questioning, designed to elicit inferential or abstract thinking, is defined as questioning that employs the higher levels of Bloom’s taxonomy: analyzing, evaluating, and creating (Anderson et al., 2001). In a recent literature review of SBR research with preschool-aged children (Walsh & Hodge, 2016), the authors concluded that the questioning strategies that teachers need to employ with young children are greatly dependent on the cognitive characteristics of the children. To stimulate further growth in language and literacy in children who display advanced abilities, teachers need to tailor the questions they ask during SBR to the child’s level of ability. From a theoretical perspective, if we want children to learn, then we need to provide educational experiences that fall into Vygotsky’s (1978) Zone of Proximal Development (ZPD). Vygotsky defined the ZPD as “the distance between the actual development level as determined by independent problem solving and the level of potential development as determined through problem solving under adult guidance or in collaboration with more capable peers” (p. 86). In terms of SBR, if teachers ask questions that are too easy, children may become bored and disengaged (Turner & Paris, 1995). Likewise, questions that are too difficult may also cause children to become disengaged, anxious, and/or frustrated (Turner & Paris, 1995). Optimal learning occurs when there is a match between the level of challenge and the level of the child’s ability.

To date, most of the research about SBR interventions in the early years has focused on children from disadvantaged backgrounds and those with disabilities or language delays (Walsh & Hodge, 2016). One study that did take into account individual differences in language ability was by Sénéchal, Thomas, and Monker (1995). The 4-year-old children in the study were divided into two groups based on pretest scores on the Peabody Picture Vocabulary Test (PPVT-R; Dunn & Dunn, 1981). The high word knowledge group had a mean score of 113 (age equivalence of 5 years 7 months; mean chronological age of 4 years 9 months), and the low word knowledge group had a mean of 93 (age equivalence of 4 years 2 months; mean chronological age of 4 years 9 months). The children were then randomly assigned to one of two book-reading conditions: passive listening or labeling (in which the adult asked the child questions that required the child to label pictures with targeted vocabulary). Following two readings of the book, the children were tested on the new vocabulary presented in the book. The children with higher prior language knowledge who had participated in the labeling condition were able to produce more new words than children with smaller initial vocabularies and children in the passive listening condition.

A further experiment introduced a pointing condition in which children were asked to point to the illustration of the target word. Again, the researchers found that children with higher prior vocabulary outperformed the children with lower prior vocabulary, but no significant interaction was found between reading condition and vocabulary knowledge; that is, no one reading condition was more effective for the children with high prior word knowledge. This study used only low-level forms of questioning (labeling and pointing to pictures) with the children, and it is unsurprising, therefore, that the children with high prior word knowledge outperformed the low prior word knowledge group.

Several researchers have examined the difference and effects on children of low demand and high demand questioning (e.g., Blewitt, Rump, Shealy, & Cook, 2009; Reese & Cox, 1999; van Kleeck, Vander Woude, & Hammett, 2006). Reese and Cox (1999) investigated differences between three styles of adult book reading: a describer style, a comprehender style, and a performance-orientation style. The describer style was characterized as low demand and involved labeling and describing by the adult throughout the book reading. In the comprehender style, there was a focus on higher demand questions about inference and prediction, with questions asked throughout the reading of the book. The performance-oriented style of reader read the book uninterrupted and asked high demand questions only at the beginning and end of the book. Reese and Cox found that children with higher initial vocabulary appeared to benefit more from the performance-orientated style for further vocabulary development, whereas children with high initial comprehension scores benefited more from the describer style, which was found to increase their print awareness skills. This research supports the idea that the manner in which teachers read to children needs to be matched to the abilities of the child and the learning outcomes that teachers are seeking.

Blewitt et al. (2009) investigated the effect of question demand level (high and low) and placement (interrupting and noninterrupting) on preschoolers’ vocabulary learning, using level of general vocabulary knowledge (as measured by the PPVT-III) as an independent variable. They did not find any interaction effects for the high and low demand questions. That is, children with initially high vocabulary scores did not do better in the high demand condition and children with low initial vocabulary did not do better with the low-level questions. However, in the same way as Sénéchal et al. (1995) had done, they did find a Matthew effect (Stanovich, 1986) present, whereby the children with initially higher vocabulary learned more new words than those with smaller initial vocabulary. Again, this study included no information about the responses that children gave to high and low demand questions during the questioning interventions.

In a second experiment, Blewitt et al. (2009) compared low demand questions, high demand questions, and a scaffolding condition in which low demand questions were asked first followed in later sessions by high demand questions. Again a Matthew effect was evident, with children with high initial vocabulary scores learning more new vocabulary than children with lower initial scores. They also found that children in the scaffolding-like condition, regardless of level of initial vocabulary, were able to provide better definitions for the new words learned than children in the low or high demand question conditions.

Although Blewitt et al.’s (2009) study would appear to contradict Reese and Cox’s (1999) findings about the interaction between prior language ability and the appropriate level of questioning demand, Blewitt et al. conceded that in many situations children with high initial vocabulary scores appeared to show more gains when a higher order questioning style is employed. They suggested that this may simply be because more advanced children learn more quickly, or it may be because a large initial vocabulary has a causal relationship with new vocabulary learning.

For all the studies described above, the primary focus was on the most appropriate questioning strategies for the acquisition of new vocabulary, rather than on ways to develop children’s ability to use higher level thinking skills such as analyzing, evaluating, and creating. In none of the research reviewed were there reports of the accuracy of the children’s responses or analysis of their responses for evidence of higher order thinking.

Higher Order Thinking in Early Childhood

There have been a number of approaches to teaching in the early years that have focused on the development of higher order thinking. For example, Katz, Chard, and Kogan’s (2014) application of the project approach encourages children to develop their own questions about the world around them. Likewise, the methods of the teachers from the Reggio Emilia city of Italy, which have significantly influenced early childhood provision, encourage children to question adults (Pound, 2011).

The work of Blank, Rose, and Berlin (1978a) focused on teaching thinking in the preschool years. Their model of language explored the interactions between teacher and children to investigate how language can support higher level cognitive processes. Blank et al. (1978a) proposed four levels of abstraction ranging from Matching Perception and Selective Analysis of Perception, to Reordering of Perception and Reasoning About Perception. The levels are hierarchical, with the two lower levels representing levels of thinking that require more literal thought and responses and the two higher levels requiring more abstract thought.

Blank et al.’s (1978a) scale attempts to capture the distance between the content material and the language. As the distance becomes wider, the cognitive demands placed on the child increase. At the lower levels of the scale, children’s language matches their experiences (Blank & Franklin, 1980). For example, a young child might label an item “cup.” Analysis of the experiences at the first level begins in the second level (e.g., the child might label the cup, “It’s blue.”). Alternatively, there may be a combination of two elements (e.g., “The milk is in the cup.”). As children progress to the third level (Reordering of Perception), language takes on a directive function. At this level, the child might refer to time and space (e.g., “This cup wasn’t here yesterday.”) or generalization (e.g., “You can drink out of cups and glasses.”). At the highest level of the scale, children are required to go beyond the information currently available to them and examine causal relationships. This requires children to use decontextualized language (e.g., “If you pour too much milk in the cup it might go on the floor.” or “This cup won’t break if you drop it because it’s made of plastic.”).

Cognitive Correspondence

There is conflicting research as to whether higher order questions do, indeed, elicit responses that indicate higher order thinking. Dillon (1982) referred to the extent to which a question of a particular level elicits a response of the same level as cognitive correspondence. Researchers working with high school and middle school samples have found that cognitive correspondence occurred in only about 50% of question-response cases (Dillon, 1982; S. R. Mills et al., 1980). These studies also found that the students were more likely to give a lower level response to a higher level question than they were to give a high-level response to a low-level question. Levels of cognitive correspondence were slightly higher for the lower levels of questioning. That is, a lower level question was more likely to prompt a lower level response than a higher order question was to elicit a higher order response.

At the preschool level, very little research has been conducted about the level of cognitive correspondence between question and response, perhaps because of the overriding tendency of preschool teachers to use lower order questions (Siraj-Blatchford & Manni, 2008) and a perception that young children are not developmentally capable of abstract and inferential thought (Blank et al., 1978a). One study that did investigate cognitive correspondence during SBR with young children was conducted by Zucker et al. (2010), who analyzed children’s responses in 25 preschool classrooms to determine if the level of abstraction of the questions asked resulted in more abstract responses by the children. They found statistically significant levels of cognitive correspondence at all four levels of Blank et al.’s (1978a) Levels of Abstraction. However, the findings relied on the total corpus of questions and responses and did not allow an analysis of individual children’s responses to the questions, as the method of data collection and the camera position did not allow individual children to be identified in that study. Furthermore, Zucker et al. focused on the use of an informational text, rather than narrative fiction. More recently, Tompkins, Zucker, Justice, and Binici (2013) found high levels of cognitive correspondence during teacher–child interactions in pretend play with small groups of 3- to 5-year-olds.

Gifted Children’s Responses to Questions

In terms of our knowledge of the ways in which gifted children respond to higher order questioning, Gallagher, Aschner, and Jenné’s (1967) seminal study of gifted high school students (aged 12–15 years) found that, even when teaching the gifted, teachers relied on cognitive memory (lower order) questions at least 50% of the time. In some of the classes that Gallagher et al. observed, there were no divergent or evaluative (higher order) questions asked at all. The authors also found that there was a consistently high relationship between the thought process asked for by the teacher (question type) and the thought process provided by the students (level of response). That is, if teachers asked a divergent thinking question, they were likely to receive a response that demonstrated divergent thinking. Similarly, if teachers asked a low-level question, they were more likely to elicit a low-level response. Gallagher et al. stated that the latter was not surprising as it is hard to respond to a low-level question such as “When was Lincoln shot?” with a response that displays divergent thinking.

With regard to young gifted children, previous research (Walsh, 2014) demonstrated that for some children, the level of linguistic complexity in their responses to higher order questions during SBR was greater than for lower level questions. However, in that study, the responses were analyzed using the mean length of communication units (c-units; Loban, 1976). C-units are defined as “each independent clause with its modifiers” (Loban, 1976, p. 9). Mean number of words per c-unit provides an indication of the linguistic complexity of a response, but not necessarily its cognitive complexity. Using this measure, a response could be incorrect yet still have a high score on mean length of c-unit. For example, in response to the question “Do you think it is really possible for Mr. Pockets to understand what the dogs are saying?” one child in the study responded with “Yes ’cause if he doesn’t know then he doesn’t know what they are telling him” (scored as one c-unit of 17 words, with contractions counted as two words). Although this was a nonsensical response, it still scored highly on the c-unit measure of linguistic complexity. In contrast, a high-level coherent response such as “because he’s a good man/and he knows what dogs mean/and he always listens to them” (scored as 3 c-units with a total of 18 words, giving a mean length of six words per c-unit), does not score as highly on the c-unit measure of linguistic complexity as the first response cited. However, it does indicate a higher level of thinking.

The purpose of the current study was to analyze the level of cognitive response made by gifted preschool children to higher order and lower order questions asked during SBR, with a view to examining patterns of thinking in response to different types of questions, and to determine whether there is cognitive correspondence (Dillon, 1982). Specifically the research questions were as follows:

Research Question 1: Are gifted preschool children able to respond to higher order questions?

Research Question 2: What is the relationship between higher order questions and higher order responses and lower order questions and lower order responses (cognitive correspondence)?

Research Question 3: Is there a relationship between question type and level of abstraction of child response?

Method

The data reported in the current study were originally collected for two studies examining questioning during SBR and its effects on the responses of young gifted children. The details of these studies are reported in Walsh and Hodge (2016) and Walsh and Kemp (2013). To examine children’s responses to higher and lower order questioning using finer distinctions than are possible with a measure of the length of c-units, data were coded and analyzed differently for the current research. Data have also been included for two participants excluded from the earlier studies because a stable baseline required for single-subject research was not achieved.

Participants

Eight young gifted children were selected for this study based on nomination by the directors of their preschools and subsequent testing on the Peabody Picture Vocabulary Test-4 (PPVT-4; Dunn & Dunn, 2007) and the Raven’s Coloured Progressive Matrices (CPM; Raven, 2004). Both of these tests are generally accepted as suitable measures of intellectual giftedness in young children (Karnes, Manning, Besnoy, Cukierkorn, & Houston, 2005; C. Mills, Ablard, & Brody, 1993; Porter, 2005; N. M. Robinson & Robinson, 1992). A percentile score of 98 or above on either instrument was used as the criterion for acceptance into the study, this being the percentile equated with moderate giftedness (Gross, 1993; Porter, 2005). The mean standard score of participants on the PPVT-4 was 136 (99th percentile; range = 117–146; SD = 8.3) and for the CPM 118 (range = 100–140; SD = 18). A standard score on the CPM could not be calculated for the three youngest participants as the test norms commenced at age 4. Pseudonyms have been assigned to the children to protect their identity.

The mean age of participants was 4 years and 2 months (range = 3 years 3 months to 4 years 9 months). There were equal numbers of girls and boys. All children were attending day care or preschool centers in Sydney, Australia, and spoke English as their sole language. The majority of the children came from middle-class families, based on the level of government support they received for their preschool/day care fees, with the exception of one child (Molly) whose mother was a university student and received additional financial support from the government for child care costs.

Setting

One private, one community, and one university child care center were recruited for the study. Five children attended the university-based center only. One child attended the community center in addition to the university center, and the remaining two children attended a private center. All centers reported having a child-centered, play-based philosophy underpinning their curriculum and reported using SBR as part of their daily program.

Materials

Forty-three high-quality children’s picture books were used in the study. The books were all either winners or had been shortlisted by the Children’s Book Council of Australia and contained narrative fiction suitable for younger children. A list of these books is available on request.

Development of Questions

Lower order and higher order questions were developed for each of the 43 books used in the study. Lower order questions were defined as “those that require responses either recalled directly from memory or cited explicitly in text” (Barden, 1995, p. 423). Higher order questions were “those that require more than simple recall to produce an answer” (Barden, 1995, p. 423). The lower level questions were developed using the lower levels of Bloom’s taxonomy: remembering, understanding, and applying (Krathwohl, 2002). The higher order questions were developed using Frank E. Williams’s (1972) Total Creativity Program for Individualizing and Humanizing the Learning Process. They included analogy, provocation, paradox, attribute listing, and tolerance for ambiguity. Although Williams (1993) acknowledged that his model was not specifically designed for gifted children, it is frequently cited as including strategies suitable for use with gifted children (Gross, MacLeod, Drummond, & Merrick, 2001; Maker & Schiever, 2005; VanTassel-Baska, 1994). Examples of the lower order and higher order questions used can be found in Table 1.

Table 1.

Examples of Question Types.

Level of question	Bloom’s/William’s question type	Definition	Examples from The Pocket Dogs by Margaret Wild	Level of abstraction
Lower order questions	Remembering	Recognizing or recalling knowledge from memory	Where is Mr. Pockets carrying his dogs?	Level 1 (Label objects)
	Understanding	Constructing meaning from different types of functions	Where does Mr. Pockets live?	Level 2 (Recall scene)
	Applying	Carrying out or using a procedure through executing, or implementing	Why did Mr. Pockets carry the dogs in his pockets?	Level 2 (Recall actions)
Higher order questions	Analogy	Situations of likeness; similarities between two things	How are Biff and Buff like Mr. Pockets’ children?	Level 3 (Compare similarities/differences)
	Provocative question	Inquiry to bring forth meaning; summons to discovering new knowledge	Can animals dream? Why do you think so? Why do you think not?	Level 4 (Text-to-life connection/comparison)
	Paradox	A situation opposed to common sense, self-contradictory statement or observation	Do you think it’s really possible for Mr. Pockets to understand what the dogs are saying? Why? Why not?	Level 4 (Judgment)
	Attribute listing	Inherent properties, ascribing qualities	Why would Pockets make good places for dogs to live?	Level 4 (Justify or explain a prediction, judgment, or inference)
	Tolerance for ambiguity	Pose open-ended situations which do not force closure	What could Mr. Pockets do if his coat became too hot in summer?	Level 4 (Explain conditions that cause alternative outcomes or a solution)

Procedure

All questions were scripted, and the place at which the question was to be asked was marked in the book with the question itself written on a sticky note. Children took part in one-on-one reading sessions conducted by the first author. After receiving informed consent from the children’s parents, the author sought consent from the individual children by asking them if they wanted to take part in the reading. If a child was reluctant to take part, then the reading did not proceed. Only one child refused to take part on one occasion, as it was her birthday and she did not want to miss a cake making activity.

The reading sessions took place in the mornings at the child’s center in a quiet room away from other children. During each reading session, the child was read one randomly selected book. Children attended from 15 to 35 reading sessions (M = 22.25 sessions) with one session a day on consecutive weekdays. The child was asked a selection of either lower order or higher order questions with questions being asked throughout the story reading session. In general, the questions were all of one type; that is, all the questions for the session were lower order or all the questions were higher order. All children took part in both higher order questioning sessions and lower order questioning sessions. In the case of one child, the intervention was varied slightly and a 50/50 mix of higher order and lower order questions were asked during some of the sessions. No feedback was given to participants during the story reading. If a child did not respond to a question or responded with “I don’t know,” then the reader (the first author) would repeat the question, or give a nonspecific prompt such as, “Can you have a guess?” or “What do you think?” If a child asked a question during the reading, the reader would respond with “What do you think?”

All sessions were recorded on an MP3 recorder with backups made on an iPhone. Transcriptions of questions and responses were made after each session.

Coding

All transcripts were imported into QSR International’s NVivo software program, a qualitative software program that assists researchers to manage, code, and analyze data.

Questions

All adult questions asked during the story reading sessions were coded for level of abstraction using Blank et al.’s (1978a) Four Levels of Abstraction (see Table 1 for definitions of each level of abstraction). The first two levels of Blank et al.’s levels were considered to represent Bloom’s lower order skills (remembering, understanding, and applying). The third and fourth levels were representative of Bloom’s higher order thinking skills (analyzing, evaluating, and creating). Within each level of abstraction each response was then coded for the question type as per the categories determined by Blank et al. (see Table 2).

Table 2.

Blank’s Four Levels of Abstraction.

Level of complexity	Coding examples^a	Example of teacher discourse^b	Example of child language behavior^c
1. Directly supplied information (Matching Experiences)	Label objects or print Locate/notice objects or print Identify sounds Imitate/repeat a simple sentence/word Count objects	What do you see? What is that? Did you see the snowman?	Labeling of objects or actions: “It’s a balloon.” Recalling information that has just been available: “That’s a whistle.” Expressing desire: “I want that.” Employing verbal routines that are tightly bound to attractive and/or familiar contexts: “Hi” upon someone entering the room. Imitating what someone else has said: “I see you” after mother has said it.
2. Classification (Selective Analysis of Experience)	Describe story actions/events/scene Imitate/act out story actions/scene Describe perceptual qualities of objects/print Recall information from a single page	How does it feel? What color is it?	Descriptions of single objects or actions: “He’s rolling.” Labeling of isolated attributes: “It’s round.” Integration of two or more distinct entities into a cohesive whole: “She’s pouring the milk.”
3. Reorganization (Reordering the Experience)	Draw an inference from something not explicitly stated in text Explain/infer character’s point of view/cognition/feelings Evaluate or judge nonperceptual qualities of objects/ideas/text as a whole Compare similarities/differences of objects, characters, or print Summarize/synthesize information across more than one page Formulate a generalization about events Draw text-to-life connection/comparison	Is an egg really round? What could we put around the snowman’s neck?	Involves concepts of time and sequence: “He came here before we went to school.” Concepts of exclusion: “All of these are red except this one.” Metalinguistic skills: “I know what W means; it means double U.” Generalization: “The knife and the scissors both cut.”
4. Abstraction and inference (Reasoning About the Experience)	Predict/hypothesize about subsequent events/conditions Define a word’s meaning Define the function/purpose of an object or print unit Explain conditions that cause alternate outcomes or a solution Identify causes of an occurrence/event Identify direct or indirect effects Distinguish between fact and fiction Provide factual background information beyond the text	What would make the snowman melt? What happens to water when you put it outside and it’s cold? What else could I use this for?	Predictions: “It’s gonna break if you drop it.” Formulating solutions: “If we put two of them together it’s not going to be long enough.” Offering explanations: “The baby can’t lift that ’cause it’s too heavy.” Providing justifications: “I could tell it would break because it was bending.”

From Zucker, Justice, Piasta, and Kaderavek (2010).

Adapted from Tompkins, Zucker, Justice, and Binici (2013).

From Blank and Franklin (1980).

Responses

All child responses to questions were coded for their level of appropriateness, ranging from fully acceptable to adequate to inadequate, and were assigned a score (see Table 3 for definitions and scores). All fully acceptable and adequate responses were then coded for their level of abstraction using Blank, Rose, and Berlin’s (1978b) Four Levels of Abstraction. Codes and definitions are listed in Table 2. When a child responded with more than one answer (e.g., a child might respond initially with “I don’t know” and then, after a prompt, reply with a fully adequate response), the highest level response to the question was coded.

Table 3.

Assessment of Child’s Response.

Coding of response	Score	Rule for coding
Fully adequate	3	Answer fully meets the demands of the task.
Acceptable
Imprecise	2	Answer valid but is vague or poorly formulated.
Oblique		Answer is not directed to the focus of the problem.
Extraneous		Answer includes extraneous or irrelevant information.
Ambiguous	1	It is not possible to determine if the answer is adequate or inadequate.
Inadequate
Invalid	0	Answer shows an understanding of the question, but the answer is incorrect.
Association to Material		Answer indicates no understanding of the question, but it is focused on the material.
Irrelevant		Answer shows no understanding of the question or the material.Answer is an imitation of all or part of the adult’s words or actions.Answer is denial of the problem stated.
I don’t know		Child states that he or she cannot answer.
No response		Child offers no verbal response.

Source. Blank, Rose, and Berlin (1978a). Reprinted with permission of the author.

Note. Blank et al. (1978a) stated that there is a difference between “I don’t know” and “No response” as “I don’t know” suggests a willingness to continue to engage with the exchange, whereas no response is indicative of a withdrawal from the dialogue.

Reliability

Interrater agreement

A final-year undergraduate Early Childhood Education student was recruited to analyze the accuracy of the transcripts prepared by the researcher. The independent rater was given 34 transcripts (19%) and the original recordings made of the sessions with the children. The rater was asked to note any discrepancies between the transcripts and the audio recordings. Percentage agreement between the researcher and the independent rater was calculated using the total agreement approach (Kennedy, 2005): dividing the smaller total by the larger total and multiplying by 100. The interrater reliability for the transcription of the sessions was calculated to be 97%.

A graduate student in early childhood was given 59 transcripts (33%) and coding instructions for level of question, question type, level of child response, and level of appropriateness of child response. Percentage of agreement was again calculated using the total agreement approach (Kennedy, 2005). The percentages of interrater agreement were calculated to be 84% for level of question, 92% for question type, 81% for level of child response, and 83% for level of appropriateness of child response.

Results

Number and Duration of SBR Sessions

The children took part in a total of 177 story reading sessions. These sessions lasted a mean length of 6.68 minutes (range = 2.68–12.00 minutes).

Questions

During the reading sessions, 997 questions were asked. Of these, 348 were lower order (Levels 1 and 2) and 649 were higher order (Levels 3 and 4) with an average of 5.6 questions per story reading session.

Gifted Children’s Responses to Higher Order Questions

To determine whether the gifted children in the sample were capable of answering the higher order questions appropriately, their responses were first coded for level of appropriateness using Blank et al.’s (1978a) Assessment of Child’s Response (see Table 3). There were 290 fully adequate or acceptable responses to lower order (Levels 1 and 2) questions and 403 to higher order (Levels 3 and 4) questions. This means that 83% of lower order questions and 62% of higher order questions received a fully adequate or acceptable response. The number of responses at each level of questioning is reported in Table 4. For Level 1 questions, 79% of responses were adequate or acceptable; at Level 2, 84% were adequate or acceptable; at Level 3, 64% were adequate or acceptable; and at Level 4, 58% were adequate or acceptable. Of the 403 adequate responses to questions asked at Levels 3 and 4, 354 (88%) were higher order (see Table 5).

Table 4.

Frequency of Child’s Adequate and Inadequate Responses at Each Level of Blank et al.’s Scale, by Level of Question.

Coding of response	Level of question
Coding of response	Level 1	Level 2	Level 3	Level 4
Fully adequate	27	202	198	81
Acceptable
Imprecise	0	54	51	25
Oblique	0	4	22	15
Extraneous	0	3	4	7
Ambiguous	0	1	10	14
Inadequate
Invalid	7	17	39	22
Association to material	0	7	23	11
Irrelevant	0	7	27	16
I don’t know	0	16	49	27
No response	0	3	5	3
Total number of Questions	34	314	428	221

Table 5.

Number (and Percentage Within Teacher Question) of Appropriate Child Responses at Each Level.

	Child response
	Level 1	Level 2	Level 3	Level 4
Teacher question
Level 1	21 (77.8)	5 (18.5)	0 (0)	1 (3.7)
Level 2	24 (9.1)	174 (66.2)	52 (19.8)	13 (4.9)
Level 3	5 (1.8)	37 (13.5)	158 (57.5)	75 (27.3)
Level 4	1 (0.8)	6 (4.7)	38 (29.7)	83 (64.8)

Mean scores for each level of questioning were calculated using the criteria outlined in Table 3. Three points were awarded for fully adequate answers, 2 points for acceptable answers, 1 point for ambiguous answers, and no points for inadequate answers with the highest mean score being 3 and the lowest, 0. The mean scores and standard deviations for appropriateness of response at each level were 2.45 (SD = 0.6991) at Level 1, 2.26 (SD = 0.5032) at Level 2, 1.71 (SD = 0.395) at Level 3, and 1.55 (SD = 0.464) at Level 4. Blank et al. (1978a) set a criterion of 2.0 or above as reflecting mastery of a level. By this criterion, 87.5% of the children in the sample showed mastery at Level 1, 75% at Level 2, and 37.5% at Levels 3 and 4.

In 48% of cases, it was found that children gave acceptable or fully adequate responses after initially giving the response “I don’t know.” In 56% of these cases, the child had been prompted to answer the question. In the other 44% of cases, the child’s initial response of “I don’t know” was followed by an unprompted more complete answer.

Cognitive Correspondence

Generalized linear mixed models using SPSS version 22.0 were run to assess cognitive correspondence, the extent to which teacher questions of differing levels corresponded with corresponding levels of child response. Both teacher question and child response were treated as ordinal variables, with levels ranging from one to four for both questions and responses. Child response was assumed to have an ordinal multinomial distribution, and the cumulative complementary log-log link function was used. The more complex mixed model was required rather than a simpler ordinal logistic regression, as one of the assumptions of ordinal logistic regression is independence of observations. In the context of the current analysis, this would mean that each teacher question/child response pair was independent. Given that each child responded to many different questions, this assumption is violated. The generalized linear mixed model was therefore run predicting child response from teacher question, with questions and responses nested within child. In other words, question/response is at Level 1, with child at Level 2, with multiple questions/responses present for each child.

Two analyses were run. The first analysis included child sex and teacher question as fixed factors, and book as a random factor. There were no effects of child sex or book (p = .520 and p = .652, respectively), so these two predictors were excluded from the second analysis (details of the first analysis available from the authors on request). The second analysis was identical to the first but had only teacher question as a fixed factor. The intraclass correlation for this analysis was .04, meaning 4% of the variance of the data set was due to the clustering within children. It could be argued that this very small proportion of variance being accounted for by the clustering renders the generalized linear mixed model unnecessary, so the same analysis was rerun through an ordinal logistic regression. The overall results of the two analyses are highly similar, and the generalized linear mixed model analysis is reported here. There was a significant effect of teacher question, F(3, 687) = 88.64, p < .0005. Pairwise comparisons were carried out comparing each teacher question with all other teacher questions. Significance levels were Bonferroni adjusted for multiple comparisons, meaning all obtained p values are compared with an alpha of .008. Parameter estimates and significance levels for each comparison can be found in Table 6.

Table 6.

Pairwise Comparisons of Teacher Questions Predicting to Child Response.

Comparison	B	SE	t	p
Level 1 vs. Level 2	−1.32	0.27	−4.84	<.0005
Level 1 vs. Level 3	−2.62	0.28	−9.31	<.0005
Level 1 vs. Level 4	−3.65	0.31	−11.65	<.0005
Level 2 vs. level 3	−1.31	0.27	−4.84	<.0005
Level 2 vs. Level 4	−2.33	0.18	−13.19	<.0005
Level 3 vs. Level 4	−1.03	0.17	−6.08	<.0005

Note. All p significant at <.008.

As seen in Table 6, all pairwise comparisons were significant, meaning that increased levels of teacher question were associated with increased levels of child response. More specifically, a Level 1 question was likely to generate a Level 1 response; a Level 2 question, a Level 2 response; a Level 3 question, a Level 3 response; and a Level 4 question, a Level 4 response (see Table 5).

Question Types

The relationship between question types within each level of teacher question and child response was analyzed with generalized linear mixed model analyses in SPSS, predicting child response (ordinal multinomial distribution as above) from question type, to obtain significance of differences between question types. At Levels 1, 2, and 4, no significant differences were found between the question types; no one type of question at that level of abstraction appeared to be more effective than another at eliciting responses of certain levels than others (p = .83, p = .07, and p = .40 for Levels 1, 2, and 4, respectively). Significant associations were found, however, between question type and child response for Level 3: F(6, 266) = 5.73, p < .0005. See Table 7 for within-question type proportions of child responses at Level 3. As seen in Table 7, questions requiring children to summarize/synthesize information across more than one page appeared to yield the lowest level responses, whereas a request to evaluate or judge nonperceptual qualities of objects/ideas/text as a whole appeared to yield the highest level responses. Post hoc comparisons were therefore carried out within the generalized linear mixed model, comparing each question type to both (a) summarize/synthesize questions and (b) evaluate or judge nonperceptual qualities of objects/ideas/text as a whole questions. Comparisons are Bonferroni adjusted for multiple comparisons, meaning all obtained p values are compared with an alpha of .0042 (12 comparisons). Table 8 gives parameter estimates and significance levels for each comparison.

Table 7.

Proportions of Question Type of Child Response at Level 3.

Question type	Level of child response
Question type	1	2	3	4
Draw text-to-life connection	0	.06	.71	.24
Formulate generalization	0	0	.67	.33
Draw inference	.01	.16	.49	.34
Compare similarities/differences	.03	.07	.83	.07
Summarize/synthesize	0	.67	.33	0
Evaluate or judge nonperceptual qualities	0	0	.45	.55
Explain/infer character POV	.03	.20	.46	.31

POV = Point of view.

Table 8.

Pairwise Comparisons Between Question Types Within Level 3 Teacher Questions.

Reference question	Comparison question	B	SE	T	p
Summarize/synthesize	Draw text-to-life connection	1.890	.6224	3.037	.003*
	Formulate generalization	2.472	.9435	2.620	.009
	Draw inference	2.128	.5614	3.790	<.0005*
	Compare similarities/differences	1.512	.5567	2.716	.007
	Evaluate or judge nonperceptual^a	2.921	.6178	4.729	<.0005*
	Explain/infer character POV	2.059	.5640	3.651	<.0005*
Evaluate or judge nonperceptual	Draw text-to-life connection	−1.031	.4172	−2.472	.014
	Formulate generalization	−.449	.8232	−.545	.586
	Compare similarities/differences	−1.409	.3240	−4.348	<.0005*
	Draw inference	−.793	.3203	−2.477	.014
	Explain/infer character POV	−.862	.3235	−2.664	.008

The corresponding comparison is not repeated in the bottom half of the table.

Comparisons which are significant at p < .0042.

POV = Point of view.

Child responses to higher order questions

Additional information about child response to teacher questions comes from examination of question–response pairs with a focus on the higher order questions that appeared to be most effective in eliciting from children a higher order response and the questions that children found difficult to answer. The results presented here focus on four higher order question types: evaluate or judge nonperceptual qualities of objects/ideas/text as a whole; compare similarities/differences of objects, characters, or print; predict/hypothesize about subsequent events/conditions; and provide factual background information beyond the text. These have been selected as they provide further insight into some of the successful and less successful strategies to elicit higher order thinking.

Evaluate or judge nonperceptual qualities of objects/ideas/text as a whole

At Level 3, significant differences were found for those questions in which children were asked to evaluate or judge nonperceptual qualities of objects/ideas/text as a whole. This category included higher order questions such as “What does a farm need to be a farm?” “Which do you think would be better to have, a baby or a pony? Why?” and “Which do you prefer, night time or day time animals? Why?” All of these questions required children to give extended reasons for their choices or generalizations. The complexity of some of the answers received revealed higher level thinking and extensive knowledge. For example, in response to the first question, “What does a farm need to be called a farm?” one child gave the following response:

You need that (points to windmill) to make power and the farm is really big. It’s got lots of trees and grass and a light. Tractor and a lawn mower. And you need farm animals. So all farms have sheep dogs. I don’t think all farms have cats. I once milked a cow. I did it a long, long, long, long time ago. (Lionel) [Fully adequate: Level 4 response]

Lionel relied to some extent on the pictures from the book to identify components of a farm (e.g., windmill, trees, grass, a light) but then was also able to question whether certain elements were essential, such as the inclusion of cats. He then offered a text-to-real-life connection by linking his knowledge of farms to his experience of milking a cow.

In the following example, Molly used her knowledge of the habits of nocturnal animals to provide a justification of her preference for diurnal animals:

Researcher:

These are nighttime animals (points to the book). Which do you prefer, nighttime or daytime animals?

Molly:

Daytime animals.

Researcher:

Why?

Molly:

’Cause they’re just more nicer. They don’t hunt around like nighttime animals. Look (points to picture of a group of owls), they’re making a collection. [Fully adequate: Level 4 response]

Compare similarities/differences of objects, characters, or print

Questions that required the children to identify similarities and differences were not as effective as some other forms of Level 3 questioning. On further examination, it was determined that these questions fell into two separate categories: those that requested identification of similarities and those that requested identification of differences. In some cases, it appeared that the children were more able to answer questions that required them to describe differences than those that asked them how two things were alike. Consider the following examples from the book Daddy’s Having a Horse (Shanahan & Quay, 2005) in which a father pretends he is pregnant with a horse while the mother is expecting a baby:

Researcher:

How are babies like horses?

Lionel:

I don’t know. [I don’t know]

Researcher:

How are they different?

Lionel:

’Cause a horse goes neigh and babies go gaga and horses have hooves and babies have feet. [Fully adequate: Level 3 response]

Lionel was unsure of how babies and horses are alike but quite confident in generating ways in which they are different. In the following example, Molly demonstrated the same difficulty:

Researcher:

How is the horse like the baby?

Molly:

Umm . . . No. [Inadequate: invalid]

Researcher:

How is the horse different to a baby?

Molly:

The pony . . . umm the baby doesn’t have a tail and the pony doesn’t have that (points to picture of the baby) and . . . that’s something that I know. [Fully Adequate: Level 3 response]

On other occasions, children appeared to confuse the terms “alike” and “different,” as in the following example.

Researcher:

How are the Bunyip and wallaby alike?

Molly:

That one’s got feathers and that one doesn’t. That’s why they’re different. ’Cause that one’s got feathers and that one’s got feathers. [Fully Adequate: Level 3 response]

In this case, the child corrects herself and gives a fully adequate response after initially giving an incorrect response.

Predict/hypothesize about subsequent events/conditions

Prediction questions, in which a child is asked to hypothesize about an upcoming story event or action, are one of the most common forms of higher order questioning employed by early childhood teachers (Hindman, Connor, Jewkes, & Morrison, 2008). However, as Walsh and Hodge (2016) highlighted, if a child has already read a story, then a prediction question such as, “What do you think is going to happen next?” becomes a simple, lower order recall question. Although prediction questions were not found to yield significantly more higher order responses than other Level 4 questions, 68% of prediction questions did elicit a Level 4 response. There was variation across the spread of answers, perhaps because of the variability of answers. For example, some prediction questions were very simple (e.g., “What do you think might be in the pirate chest?”). One child’s response was “Gold” (Lionel) [Fully Adequate: Level 3 response]. Other prediction questions resulted in much more lengthy explanations that demonstrated higher order thinking. For example, in the book Mr Chicken Goes to Paris (Hobbs, 2010), the children were asked to predict how the story would have been different if Mr Chicken had been a mouse. Alicia responded, “He would have fit on the plane” [Fully Adequate: Level 4 response]. Lionel replied, “It would be good because the mouse could just fit in. He could just go through the window” [Fully Adequate: Level 4 response]. Molly answered, “He would have been a little bit scuttly [sic] and he might bite” [Fully Adequate: Level 4 response]. All children were able to confidently predict changes that would occur in the story based on a change in the main character.

Provide factual background information beyond the text

Some questions required prior knowledge that the children did not possess. For example, in the book Fearless (Thompson & Davis, 2009), the mother places newspapers on the kitchen floor when she is house training the dog. Four of the children were asked, “Why do you think Mum left newspapers on the floor at night?” The children responded with: “’Cause he [sic] forgot to put them back” (Jeremy) [Inadequate: Invalid]; “So the dog didn’t get them?” (David) [Inadequate: Invalid]; “Because she didn’t want Fearless [the dog] to get them” (Anna) [Inadequate: Invalid]; “Because . . . because . . . because . . . ummm . . . because she wants to read them after sleeping” (Adam) [Inadequate: Invalid]. These are all logical answers, but they clearly demonstrate that the children did not possess the life experience to answer this question. In a situation where children were able to receive feedback, this misunderstanding could have been eliminated.

Questions that were within the children’s field of knowledge proved more successful. For example, in the book The Tall Man and the Twelve Babies (Champion, Niland, & Niland, 2010), the children were asked to hypothesize from where the tall man had obtained the 12 babies. Children’s responses included, “From his wife” (David) [Fully Adequate: Level 3 response]; “His tummy. Maybe he got them from the woman and the woman died so he’s by himself with the 12 babies” (Lionel) [Fully Adequate: Level 4 response]; “Maybe ’cause there was a girl there, but she died and she had a lot of babies. But maybe he found more babies” (Molly) [Fully Adequate: Level 4 response]. All children were confident in generating ideas as to the origin of the 12 babies.

Responding to Lower Order Questions With Higher Order Thinking

Only one instance was found in the entire data set where a child gave a Level 4 response to a Level 1 question. At Level 2, approximately 24.7% of questions yielded a higher order response. The Level 2 question type most likely to elicit a higher order response was describe perceptual qualities of objects/print. This type of question, such as “What does the bird look like?” yielded a level of response that was marginally higher than those in which a child was asked to recall actions/events/scene (e.g., “What did the bird do?”). This may be because the children tended to repeat, sometimes verbatim, the text of the book when responding to recall questions. For example, one child in the study was asked a recall question during a reading of Wilfrid Gordon McDonald Partridge (Fox, 1984): “Why does he (Wilfrid) like Miss Nancy?” The text in the book reads, “But his favourite person of all was Miss Nancy Alison Delacourt Cooper because she had four names just as he did.” The child’s response was, “Maybe it’s because she has four names just like him” (Lionel) [Fully Adequate: Level 2 response].

In contrast, when asked to describe the perceptual qualities of objects, the children were able to demonstrate higher levels of thinking. For example, children were asked to describe the house of one of the book characters. One child responded with “like a castle” (Alicia) [Fully Adequate: Level 3], another with “a ghost house” (Adam) [Fully Adequate: Level 3], answers that demonstrate an ability to generate similes (a higher order thinking skill), rather than simply rely on perceptual qualities such as big, brown, or messy.

Discussion

Research Questions

Are young gifted children able to respond to higher order questions?

The eight young gifted children who took part in the study demonstrated that they were able to respond to higher order (Levels 3 and 4) questions. Unsurprisingly, as the questions became more cognitively challenging, the percentage of questions answered in a fully adequate or acceptable manner decreased. This mirrors the findings of Blank et al. (1978a) that, as the level of abstraction in questions increases, children’s ability to answer decreases. Further to this, Blank et al. (1978a) found that questions at Levels 3 and 4 were too difficult for the 3- and 4-year-old children in their sample. The gifted children in the current study significantly outperformed Blank et al.’s (1978a) cohort. Given that Blank et al. (1978a) had already identified a relationship between IQ and performance on their discourse test—the Preschool Language Assessment Instrument (PLAI; Blank et al., 1978b) on which the coding system here is based—this was an expected result.

What is the relationship between higher order questions and higher order responses and lower order questions and lower order responses (cognitive correspondence)?

As both Zucker et al. (2010) and Tompkins et al. (2013) found, there was a significant level of cognitive correspondence at all levels of abstraction. That is, a Level 3 question was significantly more likely to elicit a Level 3 response, than a response that was higher or lower than Level 3. These results support earlier research in SBR contexts (e.g., Zucker et al., 2010) and in pretend play (e.g., Tompkins et al., 2013) and contribute further to the knowledge base by confirming the finding with a sample of young intellectually gifted children.

It could be reasonably expected that young gifted children would demonstrate higher levels of cognitive correspondence than children of average ability. When compared with the results obtained by Zucker et al. (2010), it would appear that the children in the current study gave more higher order (Levels 3 and 4) responses than children in the Zucker et al. study, as well as a greater percentage of appropriate responses.

Is there a relationship between question type and level of abstraction of child response?

In the present study, the effectiveness of a question type was determined by whether a child could give an acceptable answer to the question (level of appropriateness of response) and whether that answer demonstrated use of higher order thinking skills (level of response). Clearly the questions asked at the higher levels of thinking were more difficult than the lower order questions and therefore it is unsurprising that the lower level questions achieved a higher level of appropriateness. However, when examining the level of child response, it is apparent that higher order questions are more likely to produce higher order responses than lower order questions do.

With regard to the most effective types of higher order questions, the questions that required children to evaluate or judge nonperceptual qualities of objects/ideas/text as a whole were significantly more effective in terms of children generating higher order responses. It appeared that the children in this study were very comfortable with making evaluative judgments.

It is probably incorrect in some instances to label these questions as the “least effective,” in that questions that do not readily receive a higher order response may provide opportunities for teaching. In the case of the compare similarities and differences questions, it would appear that some of the children had simply not grasped the concept of similarity, hence their confusion around the word “alike.” Clark (2013) suggested that the ability to see difference is easier than the ability to see connection, as seeing connection requires a more creative and intuitive style of thinking. This type of thinking, and therefore these types of questions, may be a useful teaching technique for young gifted children to challenge them to expand their creative thinking.

Some questions were clearly less effective at facilitating higher order thinking, such as those that required children to summarize/synthesize information across more than one page. These types of question are closer in expectations to the Level 2 recall questions, with the main difference being that children had to remember what had occurred over a longer space of time in the story. Although Blank et al. (1978a) viewed these questions as requiring a higher level of abstract thinking, because the distance between the text and the child was greater, it could be argued that these questions are in fact lower order in nature. This would especially be the case with young gifted children where their exceptional memory is a commonly identified trait (Porter, 2005). Therefore, questions that rely solely on memory are likely to be easily answered and not particularly challenging.

Difficulty in Responding

All children experienced questions that they could not, or would not, attempt to answer. At some stages in the reading sessions, it seemed that “I don’t know” was almost an automatic response. It may be the case that these were the first challenging questions that these children had been asked and therefore they were uncomfortable with answering them. Other possible explanations are that these gifted children thought an instant answer was required (and were used to giving quick answers that were praised for being correct), or used “I don’t know” to give themselves time to think.

Interestingly, Blank et al.’s (1978a) research indicated that the middle-class children in their sample with higher IQs were more likely to respond with “I don’t know” than to give an answer that fell in one of the three inadequate categories (invalid, associated, or irrelevant). They hypothesized that this may be because these children were overly concerned with being correct or were inhibited about sharing their thoughts on the problem when unsure of the answer. This desire to be correct may link to the perfectionist tendencies that have been observed widely in young gifted children (Porter, 2005; Speirs Neumeister, 2015).

Implications for Practice

Young gifted children are capable of answering higher order questions and giving responses that demonstrate both a facility with language and complex thinking. However, teachers should not expect young gifted children to demonstrate their abilities spontaneously regardless of the level of questioning. It is important that teachers are aware that asking lower level questions will generally yield lower level responses and that higher order questions should be asked to give young gifted children the opportunity to both demonstrate and further develop their higher order thinking skills. Given that previous research has suggested that early childhood teacher tend to miss opportunities for higher order thinking questions (Siraj-Blatchford & Manni, 2008), professional development in this area is an urgent need.

Teachers could use some of the effective question types suggested in this article with young gifted children: evaluate or judge nonperceptual qualities of objects/ideas/text as a whole; compare similarities/differences of objects, characters, or print; predict/hypothesize about subsequent events/conditions; and provide factual background information beyond the text. Some of these question types may be novel in a preschool setting and on face value may appear excessively challenging for the children. However, teachers should persevere with higher order questioning, remembering that children may initially respond with “I don’t know” when first presented with a challenging question. Encouraging young gifted children to articulate their thinking processes and take intellectual risks is an important part of meeting the needs of these children (Porter, 2005). For children who are initially reluctant to answer, the teacher should focus on the child’s areas of strength and interest, and ask questions in a way that initiates a dialogue with a child, where the adult is demonstrating a genuine interest in the child’s responses, rather than appearing to “test” the child’s knowledge.

Teachers should model thinking about a question, especially those higher order questions to which there is no one correct answer. Asking questions, the answer to which the teacher is herself unsure of, models the way in which adults approach unfamiliar problems, as well as attempting to dispel the notion that adults “know everything.”

The propensity for children to respond by saying “I don’t know” highlights the importance of teachers’ awareness of concepts of “wait time” and their ability to use feedback to draw out answers from children. If teachers accept “I don’t know” without a prompt or allowance of more time, the child, in some cases, may view this as an acceptable response. Previous research in gifted education has demonstrated that children as young as 3 years of age are capable of modifying their intellectual behavior to conform with social norms (Gross, 1999). Therefore, if expectations in the classroom are low, a gifted child may well perform to those expectations.

The success of questions that relied on knowledge that the children already possessed highlights the need to be child focused in questioning. It also means that there are times during reading that questioning for comprehension is required to ensure that the children understand the concepts. An example is asking children if they know about house training a family pet, in the case of the example of the newspapers on the floor. In this case, an explanation of the actions of the mother in the story would have assisted the children to understand what was happening.

Limitations and Future Research

Although the number of story reading sessions was high, the number of participants represents a relatively small sample and therefore the results of the study have limited generalizability. The children, with the exception of one, demonstrated particular strengths in the area of receptive language and would therefore be expected to have more developed language than their age peers. In this way, the results are not necessarily representative of all young gifted children.

The data reported here were originally collected for two other studies—using a different methodology—and the questions were not initially developed using Blank et al.’s (1978a) Levels of Abstraction, although lower order and higher order questions were prepared. Consequently, the number of questions asked at each level of Blank et al.’s typology was inconsistent. For example, in the current study, there were only 34 questions asked at Level 1, whereas at Level 4, there were 221. Future research might look at a more even spread of questions or replicating the 70:30 lower order to higher order ratio reported in the literature (van Kleeck, Gillam, Hamilton, & McGrath, 1997).

As the original data were collected for two single-subject studies, there is no control or comparison group of typically developing children and, therefore, comparisons reported here have had to be made with data collected by other researchers. Ideally, a comparison group of children whose CPM and PPVT scores did not fall in the gifted range would assist in giving a more complete picture of whether the higher order questioning intervention is suitable for all children, or just gifted children, and whether gifted children demonstrate thinking that is significantly ahead of their age cohort.

The lack of feedback and further prompting of the children’s responses makes it difficult to ascertain how children would have responded in a more naturalistic setting. Unfortunately, the artificial nature of the original experiment prevented constructive feedback being provided to the children. In a naturalistic setting, children may benefit from feedback and encouragement to continue to answer the questions. Possibly, the children who gave the response “I don’t know” then went on to give this response to all the questions for the session because the researcher had accepted the “I don’t know” response. In a normal interaction, the adult reader would be able to provide more scaffolding and modify the question to assist the child to answer.

Other issues arising from the nature of the original studies are that the researcher was initially an unfamiliar person to the children. Although the effects of a stranger reading with the child were to some degree controlled by the regularity of the sessions over a period of several weeks and by using the same reader for all sessions, it was clear from some children’s responses during the sessions that they were more comfortable with their regular teachers. Furthermore, the reading sessions were one-on-one, rather than in a group, which is the more usual mode of SBR in early childhood settings.

Conclusion

The current study demonstrates that young gifted children are capable learners who are able to successfully answer higher order thinking questions from an earlier age than would be expected from previous research. Their propensity to answer higher order questions with higher order responses highlights the usefulness of preschool teachers’ inclusion of higher order questions in SBR with 3- and 4-year-old intellectually gifted children. This simple intervention, not often used in preschool settings, has the potential to challenge and extend young gifted children’s thinking skills.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research and/or authorship of this article.

Author Biographies

Rosalind Walsh is a sessional lecturer at Macquarie University where she teaches gifted education. She also teaches at Queenwood School for Girls, an independent school located in Sydney where she has responsibility for high-potential learners. Her PhD research won several awards including a Macquarie University Vice-Chancellor’s Commendation, The John Geake Award from the Australian Association for the Education of the Gifted and Talented, and the NSW Teachers’ Guild Research Award.

Jennifer Bowes is a retired professor from the Institute of Early Childhood, Macquarie University, and founding director of the Children and Families Research Centre, Macquarie University. She has a research background in educational and developmental psychology and has led several major research studies.

Naomi Sweller is a senior lecturer in the Department of Psychology at Macquarie University. She is a developmental psychologist with a focus on cognitive development, primarily with children in the years prior to school. Her main areas of research involve the use of gesture as an instructional and communicative tool across a range of tasks, as well as concept learning by preschool-aged children.

References

Anderson

L. W.

Krathwohl

D. R.

Airasian

P. W.

Cruikshank

K. A.

Mayer

R. E.

Pintrich

P. R.

. . . Wittrock

M. C.

(Eds.). (2001). A taxonomy for learning, teaching, and assessing: A revision of Bloom’s Taxonomy of educational objectives. Boston, MA: Allyn & Bacon.

Barden

L. M.

(1995). Effective questioning and the ever-elusive higher-order question. The American Biology Teacher, 57, 423–426.

Blank

Franklin

(1980). Dialogue with preschoolers: A cognitively-based system of assessment. Applied Psycholinguistics, 1, 127–150. doi:10.1017/S0142716400000801

Blank

Rose

S. A.

Berlin

L. J.

(1978a). The language of learning: The preschool years. New York, NY: Grune & Stratton.

Blank

Rose

S. A.

Berlin

L. J.

(1978b). Preschool Language Assessment Instrument: The language of learning in practice. New York, NY: Grune & Stratton.

Blewitt

Rump

K. M.

Shealy

S. E.

Cook

S. A.

(2009). Shared book reading: When and how questions affect young children’s word learning. Journal of Educational Psychology, 101, 294–304. doi:10.1037/a0013844

Champion

T. N.

Niland

(2010). The tall man and the twelve babies. Crows Nest, Australia: Allen & Unwin.

Clark

(2013). Growing up gifted: Developing the potential of children at school and at home (8th ed.). Boston, MA: Pearson Education.

Dickinson

D. K.

Keebler

(1989). Variation in preschool teachers’ styles of reading books. Discourse Processes, 12, 353–375. doi:10.1080/01638538909544735

10.

Dickinson

D. K.

Smith

M. W.

(1994). Long-term effects of preschool teachers’ book readings on low-income children’s vocabulary and story comprehension. Reading Research Quarterly, 29, 105–122.

11.

Dillon

J. T.

(1982). Cognitive correspondence between question/statement and response. American Educational Research Journal, 19, 540–551.

12.

Dunn

L. M.

Dunn

D. M.

(2007). Peabody Picture Vocabulary Test (4th ed.). Bloomington, IN: Pearson Education.

13.

Dunn

L. M.

Dunn

L. M.

(1981). Peabody Picture Vocabulary Test–Revised. Circle Pines, MN: American Guidance Service.

14.

Fox

(1984). Wilfrid Gordon McDonald Partridge. Malvern, SA: Omnibus Books.

15.

Gallagher

J. J.

Aschner

M. J.

Jenné

(1967). Productive thinking of gifted children in classroom interaction. Washington, DC: Council for Exceptional Children, National Education Association.

16.

Gross

M. U. M.

(1993). Exceptionally gifted children. New York, NY: Routledge.

17.

Gross

M. U. M.

(1999). Small poppies: Highly gifted children in the early years. Roeper Review, 21, 207–214.

18.

Gross

M. U. M.

MacLeod

Drummond

Merrick

(2001). Gifted students in primary schools: Differentiating the curriculum. Sydney, Australia: Gifted Education Research, Resource and Information Centre.

19.

Harrison

(2003). Giftedness in early childhood (3rd ed.). Sydney, Australia: Gifted Education Research, Resource and Information Centre.

20.

Hindman

A. H.

Connor

C. M.

Jewkes

A. M.

Morrison

F. J.

(2008). Untangling the effects of shared book reading: Multiple factors and their associations with preschool literacy outcomes. Early Childhood Research Quarterly, 23, 330–350. doi:10.1016/j.ecresq.2008.01.005

21.

Hindman

A. H.

Wasik

B. A.

Erhart

A. C.

(2012). Shared book reading and Head Start preschoolers’ vocabulary learning: The role of book-related discussion and curricular connections. Early Education and Development, 23, 451–474. doi:10.1080/10409289.2010.537250

22.

Hobbs

(2010). Mr Chicken goes to Paris. Crows Nest, Australia: Allen & Unwin.

23.

Karnes

F. A.

Manning

S. J.

Besnoy

Cukierkorn

Houston

(2005). Appropriate practices for screening, identifying, and serving: Potentially gifted preschoolers. Hattiesburg: The University of Southern Mississippi.

24.

Katz

L. G.

Chard

S. C.

Kogan

(2014). Engaging children’s minds: The project approach (3rd ed.). Santa Barbara, CA: Praeger.

25.

Kennedy

C. H.

(2005). Single-case designs for educational research. Boston, MA: Allyn & Bacon.

26.

Krathwohl

D. R.

(2002). A revision of Bloom’s taxonomy: An overview. Theory Into Practice, 41, 212–218. doi:10.1207/s15430421tip4104_2

27.

Loban

(1976). Language development: Kindergarten through grade twelve. Urbana, IL: National Council of Teachers of English.

28.

Maker

C. J.

(1986). Suggested principles for gifted preschool curricula. Topics in Early Childhood Special Education, 6, 62–73. doi:10.1177/027112148600600109

29.

Maker

C. J.

Schiever

S. W.

(2005). Teaching models in education of the gifted (3rd ed.). Austin, TX: Pro-Ed.

30.

Mills

Ablard

Brody

(1993). The Raven’s Progressive Matrices: Its usefulness for identifying gifted/talented students. Roeper Review, 15, 183–186. doi:10.1080/02783199309553500

31.

Mills

S. R.

Rice

C. T.

Berliner

D. C.

Rousseau

E. W.

(1980). The correspondence between teacher questions and student answers in classroom discourse. The Journal of Experimental Education, 48, 194–204.

32.

Porter

(2005). Gifted young children (2nd ed.). Sydney, Australia: Allen & Unwin.

33.

Pound

(2011). Influencing early childhood education. Maidenhead, UK: Open University Press.

34.

Raven

(2004). Coloured Progressive Matrices and Crichton Vocabulary Scale—Manual. London, England: Pearson Education.

35.

Reese

Cox

(1999). Quality of adult book reading affects children’s emergent literacy. Developmental Psychology, 35, 20–28. doi:10.1037/0012-1649.35.1.20

36.

Robinson

Shore

B. M.

Enersen

D. L.

(2007). Best practices in gifted education: An evidence-based guide. Waco, TX: Prufrock Press.

37.

Robinson

N. M.

Robinson

H. B.

(1992). The use of standardized tests with young gifted children. In Klein

P. S.

Tannenbaum

A. J.

(Eds.), To be young and gifted (pp. 141–170). Norwood, NJ: Ablex.

38.

Sénéchal

Thomas

Monker

(1995). Individual differences in 4-year-old children’s acquisition of vocabulary during storybook reading. Journal of Educational Psychology, 87, 218–229. doi:10.1037/0022-0663.87.2.218

39.

Shanahan

Quay

(2005). Daddy’s having a horse. Sydney, Australia: Hodder Children’s Books.

40.

Siraj-Blatchford

Manni

(2008). “Would you like to tidy up now?” An analysis of adult questioning in the English foundation stage. Early Years, 28, 5–22. doi:10.1080/09575140701842213

41.

Speirs Neumeister

. (2015). Perfectionism in gifted students. In Neihart

Pfeiffer

Cross

(Eds.), The social and emotional development of gifted children: What do we know? (2nd ed., pp. 29–40). Waco, TX: Prufrock Press.

42.

Stanovich

K. E.

(1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21, 360–401.

43.

Sutherland

M. J.

(2012). Gifted and talented in the early years: Practical activities for children aged 3 to 6. London, England: Sage.

44.

Thompson

Davis

(2009). Fearless. Sydney, Australia: ABC Books.

45.

Tompkins

Zucker

T. A.

Justice

L. M.

Binici

(2013). Inferential talk during teacher–child interactions in small-group play. Early Childhood Research Quarterly, 28, 424–436. doi:10.1016/j.ecresq.2012.11.001

46.

Turner

Paris

S. G.

(1995). How literacy tasks influence children’s motivation for literacy. The Reading Teacher, 48, 662–673. doi:10.2307/20201530

47.

van Kleeck

Gillam

R. B.

Hamilton

McGrath

. (1997). The relationship between middle-class parents’ book-sharing discussion and their preschoolers’ abstract language development. Journal of Speech, Language, and Hearing Research, 40, 1261–1272. doi:10.1044/jslhr.4006.1261

48.

van Kleeck

Vander Woude

Hammett

. (2006). Fostering literal and inferential language skills in Head Start preschoolers with language impairment using scripted book-sharing discussions. American Journal of Speech-Language Pathology, 15, 85–95. doi:10.1044/1058-0360(2006/009)

49.

VanTassel-Baska

(1994). Comprehensive curriculum for gifted learners. Boston, MA: Allyn & Bacon.

50.

Vygotsky

L. S.

(1978). Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press.

51.

Walsh

R. L.

(2014). Catering for the needs of intellectually gifted children in early childhood: Development and evaluation of questioning strategies to elicit higher order thinking (Doctoral thesis). Macquarie University, Sydney, Australia. Retrieved from https://www.researchonline.mq.edu.au/vital/access/services/Download/mq:44259/SOURCE1?view=true

52.

Walsh

R. L.

Hodge

K. A.

(2016, August 5). Are we asking the right questions? An analysis of research on the effect of teachers’ questioning on children’s language during shared book reading with young children. Journal of Early Childhood Literacy. Advance online publication. doi:10.1177/1468798416659124

53.

Walsh

R. L.

Kemp

C. R.

(2013). Evaluating interventions for young gifted children using single-subject methodology: A preliminary study. Gifted Child Quarterly, 57, 110–120. doi:10.1177/0016986212466259

54.

Williams

F. E.

(1972). Encouraging creative potential (Vol. 2). Englewood Cliffs, NJ: Educational Technology Publications.

55.

Williams

F. E.

(1993). The cognitive-affective interaction model for enriching gifted students. In Renzulli

J. S.

(Ed.), Systems and models for developing programs for the gifted and talented (pp. 461–484). Sydney, Australia: Hawker Brownlow.

56.

Zucker

T. A.

Justice

L. M.

Piasta

S. B.

Kaderavek

J. N.

(2010). Preschool teachers’ literal and inferential questions and children’s responses during whole-class shared reading. Early Childhood Research Quarterly, 25, 65–83. doi:10.1016/j.ecresq.2009.07.001