Reading and Writing

, 22:85

Effects of motivational and cognitive variables on reading comprehension

Authors

    • College of Education & Human DevelopmentGeorge Mason University
  • Stephen M. Tonks
    • Northern Illinois University
  • Allan Wigfield
    • University of Maryland
  • John T. Guthrie
    • University of Maryland
Article

DOI: 10.1007/s11145-008-9133-y

Cite this article as:
Taboada, A., Tonks, S.M., Wigfield, A. et al. Read Writ (2009) 22: 85. doi:10.1007/s11145-008-9133-y
  • 1.4k Views

Abstract

The authors examined how motivational and cognitive variables predict reading comprehension, and whether each predictor variable adds unique explanatory power when statistically controlling for the others. Fourth-grade students (N = 205) completed measures of reading comprehension in September and December of the same year, and measures of background knowledge and cognitive strategy use in December. Teachers rated internal reading motivation of each student. Results from multiple regression analyses showed that motivation, background knowledge, and cognitive strategy-use made significant, independent contributions to children’s reading comprehension when the other predictor variables were controlled. Further analyses showed the same cognitive and motivational variables predicted growth over a 3-month period in reading comprehension. Possible explanations of the observed relations between motivation, cognitive variables, and reading comprehension are presented.

Keywords

Background knowledgeCognitive strategiesComprehensionComprehension growthInternal motivationQuestioning

Introduction

Reading instructional programs increasingly focus on comprehension skills as children matriculate through school. Researchers and practitioners (Alexander & Jetton, 2000; Kintsch, 1998) have acknowledged the importance of students’ reading comprehension skills to success in a variety of school subject areas as well as other achievement outcomes. Given its importance to children’s school success, researchers are investigating what predicts the growth of reading comprehension skills. Studies have shown that both motivational (Chapman & Tunmer, 1995; Guthrie, Wigfield, Metsala, & Cox, 1999; Guthrie et al., 2006) and cognitive variables (e.g., Pressley & Harris, 2006) predict reading comprehension and other achievement outcomes. However, most studies, to date, have looked either at the relation of motivation variables to reading comprehension or the relation of cognitive variables to reading comprehension. Few works have examined how both sets of variables predict reading comprehension when controlling for the other set of variables. The overall purpose of this study was to examine how both motivational and cognitive variables predict late elementary school-aged children’s reading comprehension.

Motivation researchers have discussed how motivational and cognitive processes interact, and how each affects achievement outcomes (Pintrich, 2003; Pintrich, Marx, & Boyle, 1993; Wigfield, Eccles, Schiefele, Roeser, & Davis-Kean, 2006). In particular, such research has focused on how motivation provides an activating, energizing role for cognitive processes, which in turn can impact achievement (Pintrich; Wigfield et al., 2006). For example, Wigfield et al. reviewed work showing that motivational variables such as self-efficacy and intrinsic motivation predict students’ achievement in different areas such as reading ability, math, language arts, sports, and occupational choice. However, Pintrich noted that there is little specific information in the literature about the strength of these activating processes or how they operate. For instance, it is likely, that there are multiple motivational pathways for the energization of students’ behaviors such that some students may be motivated by their self-efficacy beliefs, whereas others may activate cognitive processes through personal interests or contextual factors. Research that examines the different ways that motivation relates to various cognitive processes speaks of the need for integrated models of motivation and cognition that has been emphasized in the motivation field (Pintrich).

In the field of reading motivation, in particular, several researchers have examined the relations among motivation variables and literacy skills. For example, research has found relationships of young children’s reading self-concept (assessed as students’ perceptions of reading competence, the difficulty of reading, and their attitude towards reading) with word recognition and reading comprehension skills (Chapman & Tunmer; 1995; Chapman, Tunmer, & Prochnow, 2000). Findings showed that children who reported negative reading self-concepts performed more poorly on reading-related tasks than did children with positive reading self-concepts (Chapman et al.).

In her study with first through fourth graders, Gottfried (1990) showed that reading comprehension positively correlated with intrinsic motivation for reading. Research with gifted populations has also shown that students with exceptionally high academic intrinsic motivation performed better on various reading measures from the elementary through the high school grades (Gottfried, Cook, Gottfried, & Morris, 2005). Also, late-elementary school students’ task-mastery goals have been found to be associated with their use of active (as opposed to superficial) learning strategies in literacy tasks (Meece & Miller, 1999, 2001), and students’ intrinsic motivation has been associated with high-level, complex literacy tasks (Turner, 1995) and reading amount and text comprehension (Guthrie et al., 1999).

In addition, research has established that specific dimensions of reading motivation (such as involvement and curiosity) and reading comprehension are correlated (Baker & Wigfield, 1999; Wang & Guthrie, 2004). This research has contributed by identifying the multiple dimensions of motivation, as well as demonstrating the specificity of motivation within the domain of reading (Guthrie et al., 1999; Wigfield, Guthrie, Tonks, & Perencevich, 2004). However, little work has been done that examines simultaneously the role of both cognitive and motivational variables on reading comprehension. Further, there is even less work that addresses the role that both cognitive and motivation predictors play in the growth of reading comprehension (Guthrie et al., 2007). Given these limitations in previous literature, in this study we examine possible ways in which cognitive and motivational variables operate in relation to reading comprehension and its growth. We turn next to specific dimensions of motivation and how they relate to reading comprehension.

Dimensions of reading motivation

Achievement motivation and motivation in specific domains such as reading are construed as multidimensional phenomena (e.g., Eccles & Wigfield, 2002; Schiefele, 1999; Wang & Guthrie, 2004; Wigfield & Guthrie, 1997). Factor analysis has distinguished at least nine components of reading motivation (Baker & Wigfield, 1999; Wigfield & Guthrie, 1997): (a) interest, (b) preference for challenge, (c) involvement, (d) self-efficacy, (e) competition, (f) recognition, (g) grades, (h) social interaction, and (i) work avoidance. Furthermore, motivations that are more internal, such as interest or curiosity, preference for challenge, and involvement have been distinguished as separate constructs in structural equation modeling from more external motivations such as grades and recognition and have been found to be strongly associated with reading comprehension not only in Caucasian students, but also in minority students and other cultures (Unrau & Schlackman, 2006; Wang & Guthrie, 2004).

In this study, we focused on five related dimensions of reading motivation and argue that they constitute a construct called internal motivation for reading. These five dimensions of motivation are: (a) perceived control, (b) interest, (c) self-efficacy, (d) involvement, and (e) social collaboration. We focus on internal motivation, rather than external, because individuals who are internally motivated show greater perseverance and sustained effort in their activities (Ryan & Deci, 2000). We focus on these five dimensions because prior research has determined their contributions to reading comprehension and literacy skills. In addition, empirical evidence has shown the interrelatedness of these five dimensions. For instance, Guthrie et al. (2007) examined these constructs with fifth-graders and found that correlations among them were statistically significant at two time points in the school year, indicating that they are indeed related to each other. These moderate correlations indicate that these dimensions of motivation are independent, while still related. In view of the interrelationships among these constructs we characterize these dimensions of motivation as representing the construct of internal motivation for reading. We describe internal motivation as strongly related to intrinsic motivation because it comes from within the individual and it moves the individual to pursue an activity for its own sake rather than for external reasons (Ryan & Deci, 2000). However, we also view internal motivation for reading as slightly different from intrinsic motivation because of the presence of self-efficacy as a distinct, and well-researched independent construct that relates to intrinsic motivation but it is still separate from it (Bandura, 1997). Lastly, we find support for the cohesiveness of internal motivation for reading on the empirical evidence that has repeatedly shown relationships between the different dimensions that comprise our measure of internal motivation and reading comprehension at different ages. We discuss each of the dimensions of internal motivation next.

Perceived control

Perceived control over reading refers to students’ choices and perceptions of their own control over their reading-related activities (Guthrie et al., 2007). Skinner and Greene (2008, in press) describe perceived control as individuals’ interpretations of the control they have over their experiences and the expectations that the self can produce desired and prevent undesired outcomes. Perceived control is often operationalized in classrooms as student choice. Perceived control and choice are associated positively with achievement in reading (Skinner, Wellborn, & Connell, 1990; Sweet, Guthrie, & Ng, 1998).

Interest

Interest has been defined as a “relatively stable evaluative orientation toward a certain domain” (Schiefele, 1999, p. 258) and described as a personal investment in an activity (Alexander & Murphy, 1998). Student interest has been shown to correlate with cognitive processes such as deeper text processing of text learning when other factors such as text length, text genre, background knowledge, and text difficulty were statistically controlled. Interest has also been found to correlate more highly with deep-level learning than with surface-level learning from texts (Schiefele, 1996; Schiefele & Krapp, 1996).

Involvement

Involvement can be defined as a descriptor of internal motivation that refers to the feeling of being absorbed in reading activities and spending significant amounts of time reading. Involvement and interest are highly related but they are still separable from each other. Devotion of time to an activity or a task denotes the individual’s involvement in it. Students who are highly involved in reading seem to create the opportunities that will support long periods of sustained reading such as organizing their activities and planning for reading time (Wigfield & Guthrie, 1997).

Self-efficacy

In both, the general motivation literature and the literature on reading motivation, one central dimension is beliefs about one’s ability, or self-efficacy. Self-efficacy refers to individuals’ judgments and perceptions about whether they are capable of doing well and accomplishing a task (Bandura, 1997). Reading self-efficacy refers to individuals’ judgments or self-evaluations about their ability to do well on reading activities such as reading a book, or reading a passage (Chapman et al., 2000; Schunk & Pajares, 2002; Wigfield et al., 2006). Reading self-efficacy has been found to correlate positively with different measures of reading, such as reading comprehension (Schunk & Rice, 1993), breadth of reading and amount of reading outside of school (Wigfield & Guthrie, 1997).

Social collaboration

Social collaboration in reading has also been studied within the motivation literature (e.g., Wigfield & Guthrie, 1997). It consists of productive social interactions among learners in relation to literacy tasks such as literature circles (Almasi, 1995), or idea circles where students share conceptual ideas from different informational texts (Guthrie & McCann, 1997). Collaboration among students in reading has been correlated with dimensions of intrinsic motivation such as curiosity and reading involvement, as well as amount and breadth of reading (Wigfield & Guthrie).

Teachers’ perceptions of student motivation

We used teacher ratings of the different dimensions of students’ motivation as our indicator of motivation, creating an overall student internal motivation score from these ratings. One of the reasons we used teacher ratings (perceptions) of students’ motivation rather than student self-report was to avoid the inherent problems of social desirability of responses to self-report measures. We also wanted to build on previous research which has used teachers’ observations or teachers’ ratings of students’ behaviors to measure motivation. For example, Onatsu-Arvilommi and Nurmi (2000) showed reciprocal relations between teachers’ ratings of students’ behaviors of perseverance on task and persistence for challenging tasks and the reading skills of 6 and 7 year-olds. Further, these investigators found that teachers’ ratings of students’ motivations predicted reading skills at a later point even after earlier levels of reading skills, overall cognitive competence, and reading-related specific competence were controlled for. More recent studies have also supported the validity of teachers’ perceptions of motivation for older, later-elementary school aged children. Specifically, external observers’ ratings of student internal motivation on the constructs of perceived control (choice), interest, involvement, social collaboration, and self-efficacy correlated significantly with teacher’s ratings of students’ internal motivation on the same constructs (Guthrie et al., 2007).

Activating background knowledge, questioning, and reading comprehension

Reading comprehension is an activity that demands high cognitive resources (Duke & Pearson, 2002). Among these cognitive resources, the role of reading strategies in supporting reading comprehension has been documented extensively. Research has repeatedly indicated that strategy instruction increases text comprehension (Duke & Pearson; National Reading Panel, 2000; Palincsar & Brown, 1984; Pearson & Fielding, 1991; Pressley, 2000; Pressley, Wharton-McDonald, Mistretta-Hampston, & Echevarria, 1998) and it has also highlighted the predictive power of reading strategies on reading comprehension (see Pressley & Harris, 2006 for a review). In this study we focus on two specific cognitive reading strategies: activating background knowledge and student questioning in relation to text.

Background knowledge has been consistently identified as having a significant role in forming an organized, coherent mental representation of text (e.g., Kintsch, 1998, McNamara, 2001; Salmerón, Kintsch, & Cañas, 2006; van den Broek, Rapp, & Kendeou, 2005). Seminal studies indicated that comprehension is strongly influenced by the degree of overlap between the reader’s background knowledge and the text content (e.g., Brown, Palincsar, & Ambruster, 1984). Later cognitive research has explored more specific roles that background knowledge plays in reading comprehension such as its interaction with text coherence for traditional printed texts and for hypertexts (McNamara, 2001; McNamara, Kintsch, Songer, & Kintsch, 1996; Salmerón et al., 2006).

Student questioning is defined as self-generated questions in relation to a text, topic, or domain (Taboada & Guthrie, 2004) and has been characterized as a self-regulatory strategy that fosters reading comprehension (Collins, Brown, & Newman, 1990; Palincsar & Brown, 1984). Research on student questioning has shown that teaching students questioning strategies, such as distinguishing between good questions from poor questions (Cohen, 1983), asking main idea versus detailed questions (Dreher & Gambrell, 1985; Palincsar & Brown, 1984; Wong & Jones, 1982), or asking questions in relation to different expository text structures (Feldt, Feldt, & Kilburg, 2002) is linked to improved reading comprehension. These instructional effects of student questioning on reading comprehension have been shown in students across the age span from third grade through college (Cohen, 1983; King & Rosenshine, 1993; Nolte & Singer, 1985; Raphael & Pearson, 1985; Scardamalia & Bereiter, 1992; Singer & Donlan, 1982; Taylor & Frye, 1992). In their extensive review of instructional studies on question generation, Rosenshine, Meister, & Chapman (1996) concluded that the majority of the authors attributed the benefits of questioning on comprehension to the fact that questioning fosters active processing of text and comprehension-monitoring. More recent work has proposed that when it comes to expository texts, students’ questions enhance reading comprehension to the extent that their questions support the conceptual knowledge structure of the text (Taboada & Guthrie, 2006).

The present study

Even though questioning and background knowledge-activation have been studied repeatedly as cognitive variables in relation to reading comprehension, and, by the same token, several dimensions of motivation have been examined in relation to reading comprehension, these cognitive and motivational variables have not been studied simultaneously in relation to text comprehension. Given the prominent role of each of these variables, the aim of this study was to examine the relative predictive power of internal motivation, background knowledge activation, and student text-based questioning on the outcomes of reading comprehension and reading comprehension growth.

We expected that student internal motivation, and student use of cognitive strategies will independently contribute to variance in reading comprehension and reading comprehension growth. Our expectation was based on cognitive accounts of reading comprehension that highlight the role of cognitive processes in reading comprehension and on accounts of the significant role that motivation plays in reading comprehension. According to leading theorists, the goal of reading comprehension is to form an organized, coherent mental representation that is similar to the structure of the text that is being read (Gernsbacher, Varner, & Faust, 1990; Kintsch, 1998). The use of reading strategies such as background knowledge activation and student questioning contributes to the building of such a coherent mental text-representation. Further, extensive research has emphasized the positive effects that students’ use of cognitive reading strategies have on reading comprehension (e.g., Duke & Pearson, 2002; Pressley & Harris, 2006; Taboada & Guthrie, 2006). Similarly and as previously stated, motivation for reading has been repeatedly related to reading comprehension and other reading achievement outcomes (e.g., see Wigfield et al. 2006 for a review).

We addressed the following two research questions:
  1. 1.

    Do motivation, background knowledge, and student questioning each make significant independent contributions to the variance in reading comprehension performance?

     
  2. 2.

    Do motivation, background knowledge, and student questioning each make significant independent contributions to the variance in reading comprehension growth?

     

Method

Participants

Fourth-grade students (N = 205) from four schools in a small mid-Atlantic city school district participated with parental permission. Table 1 shows descriptive statistics for the sample. In regards to ethnicity, our sample was somewhat more diverse than the school district as a whole, where the proportions are as follows: 8% African American, 2% Asian, 87% Caucasian, 2% Hispanic, and 1% other. With regard to students’ socioeconomic status, approximately 20% qualified for free and reduced-price meals; the district-wide average was 13%.
Table 1

Demographic data of the sample

 

Students

Percent

Total N

205

100.0

Sex

  Girls

108

52.7

  Boys

97

47.3

Ethnicity

  African American

35

17.1

  Asian

8

3.9

  Caucasian

138

67.3

  Hispanic

15

7.3

  Other/Missing

9

4.4

Measures

Five measures were used in this study: (a) background knowledge, (b) student questioning, (c) multiple-text reading comprehension, (d) Gates-MacGinitie Reading Test, and (e) internal motivation. The first three measures, (a), (b), and (c) were accompanied by a researcher-designed reading packet. We administered three alternative forms of the reading packet, each with a different theme: Oceans and Forests (Form A), Ponds and Deserts (Form B), or Rivers and Grasslands (Form C). The three reading packets were parallel in content difficulty, text structure, text difficulty, length per section, number of relevant sections and distracters, and number and type of illustrations (e.g., biome versus animal illustrations). Each 75-page reading packet contained 22 sections. Reading packets contained an equal number of easy (Grades 2–3) and difficult (Grades 4–6) texts, representing nine ecological concepts and defining information on the biomes. Texts were compiled from life science trade books and they all covered the content of ecological knowledge within life science. To ensure counterbalanced administration of text packets, students within classrooms were randomly assigned one of the three reading packets so that an equal number of students within each classroom received each packet.

The two reading comprehension measures, multiple-text reading comprehension and the Gates-MacGinitie reading comprehension test were administered in September and December of the school year (Times 1 and 2). Data for the measures of background knowledge, student questioning, and internal motivation were collected in December only (Time 2) so as to examine the association of these variables with reading comprehension growth. Teachers administered assessments in their classrooms during four 60-minute periods. Each measure is described next.

Activation of background knowledge

The measure assessed students’ activation of background knowledge on a given pair of biomes (e.g., ponds and deserts, rivers and grasslands, or oceans and forests) before reading about the topic. Students wrote what they knew about plant and animal life in their assigned biome in response to a general prompt to elicit their knowledge in the topic. This was a 15-minute, open-ended writing activity. Responses were coded using a six-level rubric (see Appendix A for the complete version of the rubric). Levels in the rubric were hierarchically organized from lower to higher levels, with lower levels including minimal or inaccurate information and higher levels including more accurate information organized in relation to a set of nine, pre-defined ecological concepts (e.g., respiration, feeding, locomotion, communication, defense, reproduction etc.). For example, at Level 1 students wrote minimal statements with very few characteristics of a biome or an organism living in the biome. These statements included neither the central ecological concepts nor definitions of the biomes. In the intermediate levels (Levels 2 and 3) students included characteristics of one or more biomes, or they presented several organisms correctly classified to one or both biomes. However, at these levels definitions and ecological concepts were not always present. At higher levels (Levels 4 and 5) students included some knowledge of ecological concepts, and relationships among different organisms and their biomes. The highest level (Level 6) was characterized by background knowledge statements that were sufficiently elaborated to denote knowledge of interrelationships among several organisms and their habitats and biomes (see Appendix A for examples of each level). Interrater agreement for 26 responses on this measure was 100% adjacent and 77% exact. A third rater resolved differences.

Student questioning

Student questioning assessed students’ self-generated questions in relation to text. After browsing the reading packet for a few minutes, students had 20 minutes to write questions about their assigned biomes and the animals and plants living in them. Questions were coded based on a four-level rubric (see Appendix B, Questioning Rubric). Question levels varied in terms of the complexity and elaboration of the requested answer. Lower level questions (Level 1) required factual or yes/no answers. Level 2 questions requested information about ecological concepts, thus they elicited at least a simple explanation about a central concept. Level 3 questions were also conceptual in their requests, but were characterized by expressing some background knowledge in the question itself. The highest question level (Level 4) consisted of questions asking about relationships among ecological concepts for a given organism or for specific relationships among organisms and their biomes or habitats.

Students wrote 0–10 questions and were given a rubric score of 1–4 for each question and a score of zero if they wrote no questions. On the basis of 10 possible questions, a student’s score could range from 0–40. The mean score for each student was used for data analysis. Interrater agreement on 100 questions for 25 students was 100% for adjacent and 90% for exact coding.

Multiple-text reading comprehension

This measure of comprehension assessed knowledge built from text. In an open-ended, constructed-response task, students wrote what they knew after reading the packet and taking notes on its content. They were given 30 minutes to respond to text and express their knowledge, with two statements of encouragement after 7 and 13 minutes. Written responses were coded based on the same six-level rubric that was used for the measure of background knowledge (see Appendix A). Knowledge built from text was assessed by examining organization of information in response to key concepts and supporting facts. Thus, lower levels of reading comprehension included knowledge statements with few and non-essential characteristics of biomes and organisms living in them. Whereas, higher levels of reading comprehension included biome definitions and ecological concepts with specific supporting facts organized in a coherent statement. Interrater agreement for 20 responses was 100% for adjacent and 80% for exact coding. A third rater resolved differences.

Gates-MacGinitie reading test

Alternative forms of the Gates-MacGinitie Reading Test, Comprehension subtest (Level 4) were administered in a 50-minute period and the extended scale score was used for data analysis. The Comprehension subtest consists of fiction and non-fiction passages from various content areas for which students answer multiple choice questions. Some of the questions require answers to information that is explicitly stated in the passage, whereas others require constructing answers based on implicit information. Across-time reliability (September to December) was r (205) = .75, p < .001.

Internal motivation

The internal motivation measure used in this study consisted of five items that measured the five dimensions of internal motivation described earlier. Teachers answered five items about each student in their class. The purpose of the internal motivation measure was to assess the extent to which each student was a motivated reader within the classroom, according to the teacher’s perception. Teachers rated the students in their classrooms on the following items: (a) reads favorite topics and authors (interest); (b) thinks deeply about the content of texts (involvement); (c) is a confident reader (self-efficacy); (d) enjoys discussing books with peers (social collaboration); and (e) often reads independently (perceived control in reading). Teachers rated their students in a 20-minute session, after repeated observations of students’ behaviors and attitudes towards reading and reading activities. The response format was Not True (1) to Very True (5) and students received a score between 5 and 25. Cronbach’s alpha reliability of all items was .90 for this sample (N = 205), which indicates very high reliability.

Results

The means and standard deviations of all the variables are reported in Table 2, while Table 3 reports correlations among the variables. Note that the two reading comprehension measures, the Gates-MacGinitie (GM) and the multiple-text reading comprehension (MTC), were administered at Times 1 and 2. Data for the remaining variables were collected at Time 2. To examine our research questions, we conducted a series of multiple regression analyses, the dependent variable being reading comprehension at Time 2 (either GM or MTC) and the independent variables being motivation, background knowledge and questioning.
Table 2

Means and standard deviations of variables used in multiple regressions

Variable

M

SD

Gates-MacGinitie Time 1

486.75

47.37

Gates-MacGinitie Time 2

501.93

44.54

Mult Text Comp Time 1

2.85

1.14

Mult Text Comp Time 2

3.33

1.28

Background Knowledge

2.47

.85

Motivation

19.39

4.79

Questioning

1.46

.54

Notes: n = 205

Mult Text Comp = Multiple-Text Reading Comprehension

Table 3

Correlations among measures of reading comprehension, background knowledge, motivation and questioning

 

1

2

3

4

5

6

1. Gates-MacGinitie Time 1

     

2. Gates-MacGinitie Time 2

.75**

    

3. Mult Text Comp Time 1

.49**

.39**

   

4. Mult Text Comp Time 2

.46**

.50**

.41**

  

5. Background Knowledge

.34**

.39**

.31**

.41**

 

6. Motivation

.51**

.49**

.38**

.42**

.35**

7. Questioning

.36**

.40**

.21*

.27**

.27**

.20*

Notes: n = 205

Mult Text Comp = Multiple-Text Reading Comprehension

p < .01, two-tailed

** p < .001, two-tailed

Predictors of reading comprehension performance

Our first research question asks whether motivation, background knowledge, and student questioning accounted for significant variance in reading comprehension performance independent of one another, that is, when the other two variables were statistically controlled. To address this question, we performed six regressions (three using GM Time 2 and three using MTC Time 2) in order to enter each independent variable as the third step. All together, the three variables explained 36.3% of the variance in GM and 26.9% in MTC. When entered in the third step of the regression equation, each variable contributed a statistically significant amount of variance in both GM and MTC (Table 4). These analyses support an affirmative answer to our first research question: Each of these variables added significantly to the variance in each of two measures of reading comprehension after controlling for the other two variables in the regression equation.
Table 4

Regression statistics for motivation, background knowledge, and questioning when entered as third step in six separate multiple regressions

Dependent variable

ΔR2

Final Beta

Gates-MacGinitie Time 2

  Motivation

.118

.37***

  Background Knowledge

.027

.18**

  Questioning

.074

.28***

Multiple-Text Reading Comprehension Time 2

  Motivation

.076

.30***

  Background Knowledge

.059

.27***

  Questioning

.018

.14*

Notes: The first and second steps of each equation were the two independent variables not used in the third step (Motivation, Background Knowledge, or Questioning)

p < .05, ** p < .01, *** p < .001

Predictors of reading comprehension growth

Our second research question asks whether motivation, background knowledge and questioning explained variance in reading comprehension growth. We operationalized growth by entering Time 1 reading comprehension into the regression prior to the other three independent variables. Such a test provides an extremely strong statistical control in that a large portion of the variance in the dependent variable is explained by the variable of prior reading comprehension in the first step of the regression equation. This procedure has been used in previous research (Allen, Cipielewski, & Stanovich, 1992; Onatsu-Arvilommi & Nurmi, 2000). It is based on the assumption that when a measure of reading achievement administered at an earlier date (in this case Time 1), acts as a control for a measure of reading achievement administered at a later date (in this case Time 2), then a third variable that was associated with the later measure of reading achievement can be said to be a predictor of growth in reading comprehension.

Results indicated that GM Time 1 accounted for 56.1% of the variance of GM Time 2 and MTC Time 1 accounted for 16.8% of the variance in MTC Time 2. After entering Time 1 reading comprehension, background knowledge and questioning, motivation still added significantly to the variance in reading comprehension growth when measured with GM and with MTC (Table 5). Similarly, when background knowledge was entered last in the regression equation, it added significantly to growth in both measures of reading comprehension (Table 5). Lastly, when entered last, questioning contributed significantly to growth in GM, although its contribution to growth in MTC was weaker (Table 5).
Table 5

Regression statistics for motivation, background knowledge, and questioning when entered as fourth step in six separate multiple regressions

Dependent variable

ΔR2

Final Beta

Gates-MacGinitie Time 2a

  Motivation

.009

.11*

  Background Knowledge

.008

.10*

  Questioning

.016

.14**

Multiple-Text Reading Comprehension Time 2b

  Motivation

.040

.23**

  Background Knowledge

.041

.23**

  Questioning

.012

.12†

Notes: a The first step was Gates-MacGinitie Time 1; b The first step was Multiple-Text Reading Comprehension Time 1; The second and third steps of each regression were the two independent variables not used in the fourth step (Motivation, Background Knowledge, or Questioning)

p = .06, * p < .05, ** p < .01, *** p < .001

Discussion

In recent years, motivational researchers have called for research that helps the field understand how motivational constructs relate to various cognitive processes, in such a way that more integrated models of motivation and cognition emerge (e.g., Pintrich, 2003). The present study contributes to extant work on the relations of motivational and cognitive processes to reading comprehension by showing how motivational and cognitive variables independently predict reading comprehension. Results support the notion that even with strong statistical controls, internal motivation, as well as the cognitive variables of background knowledge and student questioning, make significant and independent contributions to variance in two separate measures of reading comprehension. In addition, each of the predictor variables contributed significantly to growth in reading comprehension with the effects of previous comprehension controlled.

Thus, this study contributes to extant literature in two main ways. First, results from this study allow examining the specific contributions of internal motivation to reading comprehension, when the contributions of two important cognitive processes or strategies are simultaneously taken into account. To our knowledge this study constitutes a first attempt in this regard. Recent investigations have delved more deeply into whether specific dimensions of reading motivation contribute to growth in reading comprehension (Guthrie et al., 2007). Findings have indicated that indeed motivational constructs such as student choice, involvement, and interest predicted reading comprehension growth after controlling for students’ initial reading comprehension. However, no other cognitive variables have been taken into account in these analyses. Thus, as noted by these authors in past investigations (e.g., Guthrie et al., 1999) there is an absence of studies measuring reading strategies independently of text comprehension itself, and measuring the simultaneous contribution of these variables and of internal motivation to reading comprehension and its growth.

We propose that it is not the predominance of cognitive processes over internal motivation or of internal motivation over cognitive processes that explain the contribution of these variables to reading comprehension. Rather, our data support the view that background knowledge, student questioning, and students’ internal motivations make independent contributions to students’ reading comprehension. We view these independent contributions as indicators of the importance of each of these variables in relation to reading comprehension. However, and in accordance with many theories of motivation (see Pintrich, 2003; Wigfield et al., 2006), we see internal motivation as the energizer of these linkages helping students to engage their cognitive processes and strategy use, which leads to growth in comprehension. We suggest that an internally motivated reader will be more devoted to reading and thus comprehend better. In other words, if internal motivation for reading is present and fostered in students, the cognitive processes of background knowledge activation and student questioning become more fluent, enhancing students’ text comprehension. Internally motivated readers have a desire to comprehend text. This desire to understand energizes the use of reading strategies by causing the reader to be metacognitive, whether it is by asking a question, forming a summary of what has been read, or activating background knowledge to build a fuller text representation.

How are the two specific reading strategies examined in our study (i.e., activation of background knowledge and student questioning) energized by a reader’s internal motivation? With respect to readers’ activation of background knowledge, it is plausible that internally motivated students are better able to remember what they are reading and better at building stronger and more stable knowledge representations. Then, with further reading, internally motivated readers may be better able to connect text to their background knowledge and continue to build fuller and richer text representations.

With respect to readers’ questioning, this is a reading strategy that by its characteristics denotes not only cognitive, but also motivational attributes of a reader. From a motivational standpoint, a reader who asks a relatively large number of high-quality questions conveys her curiosity, inquisitiveness, and interest in the topic and the text at hand. Research in student questioning has described this curiosity as the active, initiative-driven predisposition of learners who pose a substantial number of questions (e.g., Collins, Brown, & Newman, 1990; Graesser, McMahen, & Johnson, 1994) but this research has not necessarily linked these dimensions of questioning to specific motivational constructs. We believe such linkages can be made. For example, student self-generated questions express their interest in relation to the topic they are about to read; when given the opportunity to ask their own questions in relation to text students are empowered to (a) set their own goals for reading and (b) select and process certain types of information in preference to others, a characteristic central to the notion of interest (Hidi, 1990). Student questions also encompass possibilities for perceived-control and autonomy. By writing their own questions students become aware that they are not merely responding to the teacher’s or test maker’s questions, but rather they have an opportunity to decide what is of relevance in their reading and then pursue this relevant information by seeking answers to their questions. Lastly, student-generated questions can also embody opportunities for self-efficacy development, especially when students are taught to differentiate among question types or levels and are provided with opportunities to compare their current performances with past performances in generating questions and note their progress in the use of the strategy.

The second major contribution of this study to the literature rests on its instructional implications. Given that the results of this study showed that both cognitive reading strategies and internal motivation contribute independently to students’ reading comprehension and its growth, educators and practitioners need to take into account the significance that both of these practices have for reading comprehension instruction. The benefits of cognitive strategies for reading comprehension have been well established (e.g., Duke & Pearson, 2002; National Reading Panel, 2000; Pressley & Harris, 2006). However, educators need also to consider how reading strategies can be taught and fostered in the light of supporting students’ internal motivation for reading. In other words, students need opportunities to use reading strategies in a classroom context where internal motivation is equally supported through concrete practices. For example, summarizing a text or asking questions in relation to a text that is disconnected from students’ backgrounds or for which students do not have a broader context to relate the content to (e.g., completely disconnected from units that students are learning in social studies or science) will not be as successful as providing students with texts that relate to their backgrounds, or with texts for which students can make connections to ideas learned in science or social studies (Guthrie et al., 2004, 2006). If teachers can incorporate principles that support building different aspects of internal motivation for reading they most probably have higher chances of having students use reading strategies successfully, and in turn, become better comprehenders (Guthrie et al., 2004, 2006). Further, these dimensions of motivation have been discussed extensively in terms of classroom practices that can be supported and developed by teachers in classrooms across the age span (see Stipek, 1996; Perry, Turner, & Meyer, 2006). The contributions of both cognitive and motivational factors to reading comprehension and its growth, evident in the results of this study, serve to emphasize that both are equally important in the development of students’ reading comprehension and neither should be neglected in classroom instruction.

Limitations and considerations for future research

The present study has some limitations that should be acknowledged. First, because the study is correlational it does not provide information about the processes involved in the observed relations of internal motivation, cognitive processes, and reading comprehension. From this study we gleaned some information about the strength of these relations and also that motivational and cognitive variables both predict comprehension, but data from this study only allows us to hypothesize about possible explanations for the relationships among these variables. Second, only two reading strategies were used in these analyses, thus future research should examine other cognitive variables in these categories. Similarly, a composite internal motivation variable was used in these analyses. In future work it would be interesting to examine the separate dimensions of internal motivation. Finally, we studied the relations of reading motivation, cognitive processes, and reading comprehension in fourth-grade students. Future studies should examine these relations developmentally, to see when they begin to emerge and whether they get stronger as children get older.

Based on these limitations and emerging trends in the field of reading motivation we consider three avenues for future research. First, we suggest that researchers should begin studying how motivation, cognitive processes, and reading comprehension relate. This could be done through interview studies to ascertain individuals’ understandings of how their motivation relates to their cognitive effort, and reading strategies in particular. Such studies could ask students directly about their perceptions of these relations. Gaining a better understanding of the processes involved in such relations will help educators develop more effective interventions to enhance both the motivation for reading and the use of cognitive reading strategies.

Second, in this paper, we discussed ways in which motivation energizes or activates cognitive processes. Both Guthrie et al. (2004) and Pintrich (2003) suggested that cognitive processes also might influence motivation. For instance, when given an activity or task in school, students’ background knowledge with respect to that activity may activate motivational processes and beliefs, such as their self-efficacy or interest. If they know a lot about the activity, they may feel more efficacious about taking on a new activity in this area, and also, may be more interested in it. Guthrie and colleagues suggested that when the students participate in reading activities which provide strong content goals and contain rich topical content, students become more motivated to engage in and to gain knowledge from these activities. Thus, future research should examine the reciprocal ways in which cognitive and motivational processes interact.

Third, our data suggest that motivation contributed to reading comprehension independently from students’ background knowledge and their questioning in reading. This implies that the motivation effect was not attributable to these two powerful cognitive processes. Thus, as our data indicates, motivation may be an affective construct that directly influences reading comprehension. However, it remains possible that the motivation effect is mediated by a cognitive variable that was not measured in this study and that needs to be tested through the use of path analyses or other statistical procedures that were not used in the present study. For example, inferencing is a powerful memory-based process that was not measured, nor was comprehension monitoring, a metacognitive process. Either of these could mediate the effect of motivation on comprehension. Thus, although motivation appears to contribute independently from two cognitive processes (background knowledge and questioning strategy) there are additional cognitive variables that should be tested as potential mediators of the effect of motivation on reading comprehension.

Acknowledgements

The work reported herein was supported by the Interagency Educational Research Initiative (IERI) (Award #0089225) as administered by the National Science Foundation. The findings and opinions expressed here do not necessarily reflect the position or policies of the Interagency Educational Research Initiative, the National Science Foundation, or the University of Maryland. The authors of this manuscript thank Eileen Kramer and Vanessa Rutherford for their assistance in preparing this document.

Copyright information

© Springer Science+Business Media B.V. 2008