Word identification, metacognitive knowledge, motivation and reading comprehension: an Australian study of Grade 3 and 4 pupils.
Reading comprehension is an important literacy competency developed in the early years of schooling. From a self-regulated learning perspective, reading comprehension involves the interaction of cognitive, metacognitive, and motivational variables (e.g., Dignath & Buttner, 2008). Research has suggested that good reading comprehension is the result of the use of a range of these variables including word identification and decoding abilities (Chapman, Tunmer & Prochnow, 2000), knowledge of cognitive and metacognitive strategies, such as planning and self-monitoring (Pressley & Harris, 2006), and motivational aspects of learning, such as self-concept and interest (Miller & Faircloth, 2009). There is also evidence that pupils who apply cognitive and metacognitive strategies are better comprehenders (Paris, Lipson & Wixson, 1994), and that the training of such variables can lead to improved reading comprehension (Pressley, 2006). Gender may also lead to differences in reading comprehension (Logan & Johnston, 2010) The aim of this article is to explore the role that these variables play in the reading comprehension of Australian male and female pupils in Grades 3 and 4.
In the following section, we briefly summarise the major research findings concerning the state-of-the art with respect to the acquisition of reading comprehension by making reference to the role of word identification, metacognitive knowledge and motivation in reading comprehension. With respect to motivation we focus on the intrinsic motivation variables of reading self-concept and interest. We also highlight the research that discusses issues of gender difference in reading comprehension. Thereafter we outline the rationale and main goals of the present study. But we begin by defining reading comprehension, the main outcome variable in our study.
Reading comprehension refers to making meaning at the word, sentence and text level. It involves the dynamic interplay of a range of knowledge, processes and strategies (Oakhill, Cain, & Bryant, 2003). Successful reading comprehension occurs as a result of the interaction between both reader and text factors (Sweet & Snow, 2003). Of interest in this study are some of the reader variables that are brought to the reading comprehension process. One of the main variables that has been extensively studied in relation to reading comprehension is that of word identification.
A pupil's ability to read words accurately influences their reading comprehension (Jenkins, Fuchs, van den Broek, Espin, & Deno, 2003) and is a strong predictor of reading comprehension (Vellutino, Scanlon, & Tanzman, 1994). Successful word identification skills depend upon the effective utilisation of the alphabetic code and identifying words easily and rapidly. When word identification is fluent and decoding becomes accurate and automatic, cognitive resources are freed up so that the meaning of what is being read can be derived. In this way accurate word identification and decoding at a level of automaticity allow pupils to place their efforts into comprehension (Samuels, 2006). Thus successful reading comprehension occurs when readers are able to accurately and rapidly identify words and to use non-visual information such as grammatical and semantic knowledge to work out what the text says. In turn these contribute to establishing what the text means.
Metacognitive knowledge, an aspect of metacognition, is also important for successful reading comprehension. Metacognitive knowledge refers to the declarative, procedural and conditional knowledge associated with learning, for example learning to read (Pressley, 2002). Several studies have emphasised the importance of declarative and procedural metacognitive knowledge for the development of reading comprehension (see reviews Schneider & Pressley, 1997). Measures of declarative metacognitive knowledge seem particularly well-suited to predicting reading comprehension in children. For instance, Schlagmuller, Vise, and Schneider (2001) showed that metamemory as assessed by the Wurzburg Metamemory Test (Vise, Schlagmuller, & Schneider, 1998) was closely associated with memory and was a good predictor of reading comprehension in primary school children. The Wurzburg Metamemory Test (Vise, Schlagmuller, & Schneider, 1998) was used in this study.
While cognitive and metacognitive knowledge and strategies are important to reading comprehension, a student's affect also influences their comprehension. Indeed, Guthrie and Wigfield (2000) have argued that 'motivational processes are the foundation for coordinating cognitive goals and strategies in reading' (p. 408), and Wang and Guthrie (2004) found that intrinsic motivation was positively related to reading comprehension, after controlling for other variables such as extrinsic motivation and amount of reading.
Motivation in reading is comprised of a number of intrinsic and extrinsic variables (Deci, Koestner, & Ryan, 2001). As indicated earlier, in this study we focus on two variables identified with intrinsic motivation, namely self-concept and interest. With respect to the relationship with self-concept and academic achievement in general, a meta-analysis carried out by Hattie (2009) found a positive correlation of.43 between self-concept and achievement. Many researchers have examined self-concept in relationship to specific academic domains, such as reading. Reading self-concept refers to the ability to make a cognitive appraisal of one's comprehension and learning (Marsh, 1990). Chapman and Tunmer (1995) investigated the relationship between reading self-concept, letter and word identification and comprehension strategies amongst primary school pupils between 5 to 7 years of age. They found that reading self-concept correlated positively with comprehension strategies for the older pupils, while perceptions of task difficulty correlated significantly with reading strategies for the younger pupils. More recent findings suggest that pupils with lower reading self-concepts performed more poorly on reading tasks and had lower overall academic self-concept scores (Chapman, Tunmer, & Prochnow, 2000). Rider and Colmar (2006) also found a strong positive relationship between reading achievement, namely accuracy and comprehension, and reading self-concept scores for children in Grade 3.
Interest is another variable related to intrinsic motivation which was also examined in this study. Interest is defined as either a characteristic of the person or of the text (Renninger, Hidi, & Krapp, 1992). A number of researchers have argued that a student's level of interest can be activated and can vary according to both the situation and the context (e.g., Schiefele, 1991). Several studies have found that students' perceptions of interest in reading influence reading comprehension (Guthrie & Wigfield, 1999). As is the case with self-concept, these studies have indicated that interest is related to increased learning, persistence and effort. Specifically, interest in reading has been found to influence students' engagement with text, as well the quality of their comprehension (Wigfield, 1997).
Gender differences have consistently been demonstrated in the relevant literature, showing that girls are better at reading comprehension than boys. As Logan and Johnston's (2010) recent review noted this finding is evident in several studies (e.g., Mullis, Martin, Kennedy, & Foy, 2007), regardless of the type of instruction the pupils have received (Ming Chui & McBride-Chang, 2006).
Gender differences of other reading-related variables have also been found in a number of studies. For example, research related to pupils' reading attitudes and reading motivation and their association with or prediction of reading comprehension has consistently demonstrated gender differences (e.g., Wang & Guthrie, 2004), with girls scoring more highly than boys. A number of authors have argued that self-concept is determined by one's frame of reference. That is, pupils compare themselves and their performance with their local setting, namely the achievement of their peers, and possibly same-gender peers, and make comparisons related, first and foremost, to peers in their own classroom and school (Moller & Koller, 2004; Pekrun & Zirngibl, 2004). Thus this study allowed us to explore issues related to these motivational variables with respect to gender differences, and this issue of frame of reference.
Rationale for the study and research questions
The present study was conducted by a group of international researchers who are interested in the reading comprehension and the variables that impact on it. The study is situated within social-cognitive models of self-regulated learning (e.g., Winne, 2005). While there are a number of models of self-regulated learning (see Puustinen & Pulkkinen, 2001 for a review), this study fits best within models of self-regulated learning that consider self-regulated learning to be an interaction of cognitive, metacognitive, and motivational processes (e.g., Boekaerts & Corno, 2005).
The study draws on the literature summarised above that indicates that the variables of word identification, metacognitive knowledge, and motivation are related to reading comprehension. In terms of metacognitive knowledge, we specifically chose to examine the role of declarative metamemory. In the area of motivation, we chose to examine reading self-concept and interest in reading. The study not only examined the relationships among those variables associated in the literature with the development of reading comprehension, but also used teacher judgment as a measure of reading achievement. Specifically, reading comprehension was assessed by combining scores from a standardised measure of reading comprehension and a measure of teacher judgment of reading achievement. The study addressed the following research questions:
1. Are there systematic gender differences on the variables of interest?
2. Do the variables of interest correlate and what is the pattern of intercorrelations?
3. Do the variables of interest predict reading comprehension?
4. Does the causal structural model apply to the data?
We assumed that the general model (based on the regression analyses) would apply to the data, with an expectation that word identification would have an impact on reading comprehension in the sample.
The study involved a sample of 139 Australian third (M = 30, F = 31) and fourth (M = 34, F = 44) graders drawn from two small primary schools in adjoining suburbs of the city of Brisbane, Queensland, Australia. The population of School A was described as being from a 'middle' socio-economic background. The population of School B was described as being from a 'mixed' socioeconomic background. None of the children were identified as having developmental, emotional or sensory disabilities. The teachers of these pupils also participated in the study. There were four Grade 3 teachers (male N = 2, female N = 2) and four Grade 4 teachers (female N = 4). The Grade 3 teachers had between 12 and 30 years of teaching experience (Mean = 18.75, Median = 16.5) and the Grade 4 teachers had between 6 and 30 years (Mean = 21, Median = 24) of teaching experience.
A number of different instruments were used to assess performance in word identification, metacognitive knowledge, reading self-concept, interest in reading, and reading comprehension.
The Word Identification Subtest--Test 3 of the Woodcock Reading Mastery Tests-Revised (Form H, Woodcock, 1987). The subtest requires the identification of single words that appear in a list of 100 items arranged in order of difficulty. The test was administered individually and testing stopped when the pupil identified six consecutive words incorrectly. The test requires accurate word reading. The pupils' raw scores were calculated. The variable was labelled WORDID. The split-half reliability coefficients for Form H for Grade 3 is [r.sub.sh] = .97.
Index of Reading Awareness (Jacobs & Paris, 1987). This is a twenty-item test that assesses pupils' metacognitive knowledge about strategies that can be used in reading. It is a multiple-choice questionnaire that measures evaluation, planning, regulation and conditional knowledge. Five questions measure each aspect of metacognition. Pupils select one of three responses that they believe to be the best strategy in a described reading scenario. The test was administered to the class as a whole and took 10 min. Each response is scored as 0 for an inappropriate response, 1 for a partially adequate answer, or 2 for a strategic response, with a range from 0-40. A student's total score (AWARE) was calculated. Cronbach's a for the 20 items was.56. A study by McLain, Gridley and McIntosh (1991) obtained a Cronbach's alpha of.61. These authors suggested that the Scale is acceptable if used as a total score. This was done in the current study.
Wurzburg Metamemory Test (Wurzburger Metagedachtnistest) (Vise, Schlagmuller, & Schneider, 1998, rev. ed, transl). This is a test of declarative metamemory. It comprises 3 subscales. The first subscale ('Dolphin') assesses general metamemory related to person, task and strategy variables and consists of 15 items. Possible scores range from 0 to 30. The second subtest ('Seal') comprises 18 items which assesses strategies related to text processing, specifically task-related knowledge of clustering strategies for recall. Possible scores range from 0 to 36. 'Elephant' comprises 17 items and assesses knowledge of semantic categorisation strategies. Possible scores on this third subtest range from 0 to 34. (Contact the first author for a copy of the items in the subscales). The test was administered in a whole class setting and took 40 min. The range for the total score on this test (METAMEM) is 0 to 100. Cronbach's a was.74. The test-retest reliability after four months on the original version of this test was.70 (Vise, Schlagmuller, & Schneider, 1998).
Reading Self-Concept Scale (Chapman & Tunmer, 1993). This is a 30-item measure that assesses reading self-concept. It comprises three scales, namely Difficulty, Attitude and Competence that Chapman and Tunmer argue constitute three separate, but related aspects of reading self-concept. The Difficulty scale assesses perceptions of difficulty with reading such as pupils' beliefs that reading activities are hard or problematic. The Attitude scale assesses attitudes towards reading, specifically pupils' feelings towards and affinity for reading, while the Competence scale assesses perceptions of competence such as beliefs regarding ability and proficiency in reading tasks. The test took 20 minutes to administer as a whole class test. The total score (Reading Self-Concept Total: SELFCON) is expressed on a 5-point scale. Cronbach's a was .76. Other studies have also revealed the test's strong psychometric properties (Chapman & Tunmer, 1995; Rider & Colmar, 2006).
Interest in Reading Scale (Lesen Interessen Skala) (van Kraayenoord, 1996, transl) The scale comprises 10 items. Four items relate to children's attitudes towards reading such as 'I like to read books containing stories' and to texts of different types, and the other six items relate to children's habits and behaviours associated with reading, such as 'I get books for my birthday and at Christmas'. The test was administered to the whole class group and took 5 min. The children responded on a 3-point Likert-type scale (1 = Not true, 2 = Sometimes true, 3 = Always true). The range of possible scores is 10 (low) to 21 (high) (INTEREST). Cronbach's a was.59, indicating moderate but still sufficient consistency.
Tests of Reading Comprehension (TORCH) (Mossenson, Hill & Masters, 1995). The passage 'Lizards Love Eggs' (A3) was used to collect the data for 87 of the pupils in the sample. The Kuder-Richardson Reliability Coefficient, KR 20 for Test Booklet A. Passage A3. Lizards Love Eggs was [r.sub.tt] =.91. The two classes of Grade 4 pupils in School A had undertaken the Tests of Reading Comprehension (TORCH) using the passage A2. 'The Bear who Liked Hugging Trees' close to the data collection period. According to the TORCH Manual passage A2 overlaps with that of A3 and so the TORCH scores from the 52 pupils who had been tested using passage A2 were used in the study rather than retesting them. The TORCH involves the silent reading of the passage and then the completion of a cloze retelling of the passage by providing the correct answers in writing. The test items measure literal and inferential comprehension and pupils' abilities to synthesise information presented. The test was administered in whole class groups. The TORCH score was calculated for each student according to the instructions in the Manual. The variable was identified as TORCH. The Kuder-Richardson Reliability Coefficient KR 20 for Test Booklet A, Passage A2, The Bear Who liked Hugging People was [r.sub.tt] =.93 (Mossenson, Hill, & Masters, 1995). A pilot study by Burgon (1988) examining the relationship between the TORCH and the Progressive Achievement Test-Comprehension (PAT-RC; Elley & Reid, 1969) and the TORCH and teacher ratings of their pupils' reading comprehension on a 1 to 5 scale on 350 New Zealand students revealed that pupils' scores on the TORCH correlated well with the PAT-RC and with teacher ratings. The author also commented that the given that teacher ratings were based on standard classroom practice, the results point to the validity of TORCH for use in New Zealand classrooms. While it is not possible to draw inferences between practices in New Zealand and Australia, the PAT-RC and teacher judgments are also common assessment practices in Australia.
Teacher Judgment of Reading Achievement (Lehrerfragebogen zur Einschatzung der Lesefahigkeit) (Adapted from Marx, 1998, transl). Teachers were asked to rate each child on 6 subscales (TEACH 1 to TEACH 6). TEACH 1 asked teachers to make a judgement about the child's reading ability in general. For example for TEACH 1 that item stated, 'I judge the overall reading ability of the child as -, and the teacher responded with a rating from 1 = 'Far above average' to 7 = 'Far below average'. TEACH 2 asked teachers to make a judgement about the children's ability to read a new, unknown, age-appropriate text. TEACH 3 asked teachers to make a judgement about the children's ability to read words of new, unknown texts. TEACH 4 asked teachers to make a judgement about the children's ability to comprehend new, unknown texts. TEACH 5 asked teachers to make a judgement the children's ability to write words of unknown texts and TEACH 6 asked teachers to make a judgement about the children's ability to ability to comprehend following listening to new, unknown texts. Each Likert-type scale was used to assign a rating from 1 (low) to 7 (high). The mean of TEACH 1 to TEACH 6 was used to create a total score (TEACHJUDGE). Cronbach's a was. 97.
A word about the translation
In this study, the German language tests, specifically the Teacher Judgment of Reading Achievement, Wurzburg Metamemory Test, and the Interest in Reading Scale were translated into English by the first author. In some instances, alterations were also made to words or phrases to make them more culturally and contextually appropriate (e.g., names of rivers in the Wurzburg Metamemory Test were changed to the names of Australian rivers). The third and fourth authors whose first language is German and who are both very competent English speakers checked the German-English translation for accuracy and appropriateness of word selection.
Ethical clearance for the study was obtained from The University of Queensland and gatekeeper approval was provided by Education Queensland. Data was only collected from pupils for whom consent had been obtained from their parents. The first author and four research assistants collected the data. Two persons from the research team were in each classroom during test administration and the classroom teachers were not present. At the start of the test administration the purpose of the study was outlined to the class and the pupils were assured of the confidentiality of their data.
The order of presentation of the tests were as follows: Visit 1: Wurzburg Metamemory Test; Visit 2: Interest in Reading Scale, Reading Self-concept Scale, and Index of Reading Awareness; and Visit 3: TORCH. The Woodcock Word Identification Subtest was administered individually to each student in an office or empty classroom by the research assistants on a fourth visit. The first author approached the teachers individually and the purpose of the Teacher Judgment of Reading Achievement was explained. Written instructions were provided to the teachers so that the form could be completed in the teachers' own time. The completed forms were given to the school secretary from whom they were collected by the research assistants or the first author.
Prior to undertaking the analyses, for each construct used in the study (except gender), the raw scores of each subtest were aggregated and normed (separately) for each grade level in order to avoid different levels of performance. Next, these scores were converted into z-scores in order to compare the scores and across the variables.
The correlation between the two metacognitive knowledge measures (declarative metamemory: METAMEM and metacognitive knowledge about reading strategies: AWARE) in the sample was r =.35 (p <.01). These metacognitive knowledge measures were combined to create a latent variable (META) for the causal modelling. We also created a variable named total reading comprehension (READCOM) by combining the relevant z-scores on the pupils' reading comprehension tests (TORCH, r =.69, p <.01)) and the teacher judgments on the 6 subscales (TEACHJUDGE). Specifically, the z-scores of each construct (TORCH & TEACHJUDGE) were aggregated (separately for each grade level to avoid different levels of performance) and then again converted to z-scores in order to create a latent variable for the causal modelling that represented the dependent variable (READCOM).
As indicated earlier, the participants comprised pupils from Grade 3 and 4. Given that the differences between the two grades for Total reading were not significant the grade levels were combined. Statistical analyses were performed using SPSS 16.0 for Windows. Comparisons of the dependent (continuous) variables (test instrument scores) for gender were undertaken by the use of independent-samples t-tests. A value of p <.05 was considered significant. Using the effect size, Cohen's d provides an indication of the magnitude of the mean differences between the groups.
We first checked the age and gender distribution in the sample. There were no significant differences in mean ages for males and females with reference to Grade (3 and 4) (see Table 1), all ps >.05.
There was a significant difference in test scores for males (N = 64) and females (N = 75) with respect to total reading comprehension (READCOM), reading comprehension (TORCH), and reading self-concept, all ps <.04 (see Table 2). Specifically, girls outperformed the boys on all these measures.
Predictors of reading comprehension and their intercorrelations
With respect to the second and third research question we report the findings of the intercorrelations and stepwise multiple regression analyses for the sample.
A stepwise multiple regression was performed with the total reading comprehension score (READCOM) as the dependent variable and the other variables as independent variables (WORDID, META, SELFCON, INTEREST, GENDER). Preliminary analyses were conducted to ensure no assumptions were violated. With respect to multicollinearity, all the independent variables showed some relationship with the dependent variable. Additionally, all the bivariate correlations between each of the independent variables were below the cut-off point of r =.70 (r <.41). Supported by the Variance Inflation Factor (VIF) values (that measure redundancy between the explanatory variables in the model), multicollinearity was not violated, that is, < 1.31 and well below the cut-off of 10. The correlations amongst the 'plain' z-standardised variables of the sample appear in Table 3.
The word identification variable was the best predictor and explained 58% of the variance in reading comprehension. In step 2, the metacognitive knowledge variable was selected to go into the equation and explained an additional 6% of the variance. Reading self-concept was identified as the third predictor and was included in the model. Self-concept contributed to the prediction of reading comprehension and so the whole model explained 66% of the variance, F(3, 135) = 87.15, p <.0005. This pattern of results suggests that two-thirds of the variability in total reading comprehension was predicted by word identification, metacognitive knowledge and reading self-concept. In the final model, the three measures were statistically significant, with word identification recording the highest [beta] value ([beta] =.61, p <.001), followed by metacognitive knowledge ([beta] =.24, p <.001) and reading self-concept ([beta] =.17, p =.002) (see Table 4).
In sum, word identification explained 58% out of a total of 66% of the variance in total reading comprehension. Interestingly, reading self-concept added little to the prediction of reading comprehension.
In order to assess the interrelationships among predictor variables and determine the construct validity of the model, a latent variable causal modelling approach, using AMOS 16.0 was carried out. This allowed us to assess the causal influences of the various variables on reading comprehension. Preliminary analyses revealed that gender and grade did not have a reliable impact when included as exogenous variables. In addition, the relationship coherence between gender and total reading comprehension was r =.17 (p <.01) for the sample. Therefore these variables were not considered in our subsequent causal modelling. The motivational variables (reading self-concept and interest) formed a single latent construct. As reported earlier, metacognitive knowledge comprised the variables of declarative metamemory and metacognitive knowledge about reading strategies. The total reading comprehension variable comprised two measures--reading comprehension (TORCH) and teachers' judgment of reading. The correlation among these variables was r =.69 (p <.01). The results of the best-fitting causal model for the sample are shown in Figure 1.
In this sample both, word identification and metacognitive knowledge influenced reading comprehension directly, although there was a stronger path between word identification and reading comprehension than between metacognitive knowledge and reading comprehension. In addition, motivation had a direct effect on reading comprehension, as well as via metacognitive knowledge and word identification. The efficiency of the model was evaluated by several goodness-of-fit indices that summarise the discrepancy between the observed values and the expected values for the model. Overall, the various measures showed acceptable data fit for the model. The Adjusted Goodness of Fit Index (AGFI) was.87, and the Root Mean Residual (RMR) score was. 07. The Root Mean Square Error of Approximation (RMSEA) was
[FIGURE 1 OMITTED]
Finally, we examined the patterns of interrelationships among predictor variables. Word identification and metacognitive knowledge influenced reading comprehension directly. There was a strong path between word identification and reading comprehension, r =.65, with a less strong path between metacognitive knowledge and reading comprehension, r =.41. Motivation had a direct effect on reading comprehension as well as via metacognitive knowledge and word identification. Thus, word identification substantially influenced reading comprehension in the sample.
This article has reported on the results of an investigation of reading comprehension involving 139 Australian male and female pupils in Grades 3 and 4. Specifically, the pupils' word identification, metacognitive knowledge and motivation in relationship to their reading comprehension were examined.
The study found no significant differences in mean ages for males and females in this sample. Gender differences were found for total reading comprehension, reading comprehension, and reading self-concept, with girls performing better than boys on these variables.
These findings of gender differences in reading comprehension align with the results from the majority of studies that have found differences between boys and girls in reading comprehension (Logan & Johnston, 2010). With respect to gender differences in reading self-concept, this finding provides some support for the findings of the studies by Moller and Koller (2004) and Pekrun and Zirngibl (2004) who have pointed to the importance of one's frame of reference in determining one's reading self-concept.
The findings relating to the stepwise regression analyses indicated that reading comprehension was predicted strongly by word identification. In addition, the causal modelling found that the model created basically fit the data set. In the model there was a direct link between word identification and reading comprehension and between motivation and reading comprehension. Both metacognitive knowledge and word identification were moderator variables between motivation and reading comprehension. Thus motivation influenced reading comprehension, but was also mediated by both metacognitive knowledge and word identification. These results partially mirror the results of causal modelling undertaken in an earlier study of German pupils by van Kraayenoord and Schneider (1999) and the follow-up study by Roeschl-Heils, Schneider, and van Kraayenoord (2003). Specifically, in these studies the models indicated that metacognition had a direct effect on reading comprehension, and motivation had an indirect effect via word identification and metacognition. The role of word identification was more salient in this study that involved reading comprehension in the English language than was found in the German language studies.
The findings of the causal modelling revealed that word identification ability was strongly related to reading comprehension. Goswami (2008) has indicated that children have difficulty acquiring reading because phonology is complex with varying levels of consistency. This finding suggests that third and fourth Grade pupils still may need instruction (and intervention) in word identification in these years of primary school (Juel, 1988; Torgesen, 2000). Indeed with respect to implications of this finding for practice in the Australian school system there is a need for teachers to emphasise the development of the word identification abilities of their pupils in the early years of reading. Intervention programs that promote the development of word identification, reading comprehension and knowledge of strategies are especially necessary for pupils at-risk for reading failure (Tunmer & Chapman, 2002; Woolley, 2007).
Of interest is the finding that the motivational variables added little to the pupils' reading comprehension. One possible reason for the lack of impact of the motivation construct is that only two relevant indicators (i.e., reading self-concept and interest) were considered in this study. We know from other studies (e.g., Paris & Paris, 2001) that other concepts such as perceived self-efficacy might also play an important role. Another reason could be that individual differences in learning motivation are not that pronounced in third or fourth grade, but prove more important for the prediction of learning outcomes as pupils become older (Pintrich & Schrauben, 1992). Including additional motivational indicators could have improved the predictive quality of this construct and future studies could investigate the influence of other intrinsic motivation variables on reading comprehension.
There are several limitations of our study that are important to note. First, the sample is not representative of all Australian children and was limited to pupils from two schools in Brisbane. Second, the translated instruments were not piloted. Third, we caution that these results cannot be generalised beyond the particular variables related to reading comprehension that we investigated. Additional studies exploring these issues with other samples at other grade levels are required.
A strength of this study was the use of different data analysis techniques, that is both regression analyses and causal modelling. This meant that the interplay among the constructs of interest in this study could be carefully investigated.
There are several implications for research and practice. There is a clear need for effective initial teaching of reading and reading comprehension. Teachers in the early years of schooling need to ensure that all the elements of early reading acquisition are taught systematically and thoroughly. Pupils' progress should be monitored and early identification of difficulties should be followed up with early intervention (Snow, Burns, & Griffin, 1998). Education systems and schools should identify programs that will promote reading comprehension. Such programs must pay attention to the development of word-level skills as well as the development of reading comprehension (National Institute of Child Health and Human Development, 2000). In addition, such programs should be delivered in a balanced manner that integrates the teaching of word-level skills and reading comprehension strategies within meaningful opportunities for reading (Pressley, 2006; Taylor, Pearson, Garcia, Stahl, & Bauer, 2006). Other authors (e.g., Guthrie & Humenick, 2004; Taboada, Tonks, Wigfield, & Guthrie, 2009) have argued that programs in which motivation, metacognitive knowledge and strategy use are emphasised can enhance reading achievement. Such programs provide the use of interesting texts, offer pupils choice and opportunities for personal control during learning, set knowledge goals and use activities that employ cooperative learning.
To conclude, this study showed that individual differences in primary school pupils' reading comprehension can be explained by word identification, and to a lesser extent metacognitive and motivational variables. The creation and delivery of effective programs that promote successful reading comprehension for these students is essential.
Boekaerts, M.M., & Corno, L. (2005). Self-regulation in the classroom: A perspective on assessment and intervention. Applied Psychology: An International Review, 54(2), 199-231.
Burgon, J.R. (1988). Tests of Reading Comprehension (TORCH) pilot study. Wellington, New Zealand: New Zealand Council for Educational Research. Retrieved from ERIC database. (ED326855).
Chapman, J.W., & Tunmer, W.E. (1993). Reading Self-Concept Scale. Unpublished scale, Massey University, Educational Research and Development Centre, Palmerston North, New Zealand.
Chapman, J.W., & Tunmer, W.E. (1995). Development of young children's reading self concepts: An examination of emerging subcomponents and their relationship with reading achievement. Journal of Educational Psychology, 87, 154-167.
Chapman, J.W., Tunmer, W.E., & Prochnow, J.E. (2000). Early reading-related skills and performance, reading self-concept, and the development of academic self-concept: A longitudinal study. Journal of Educational Psychology, 92(4), 703-708.
Deci, E.L., Koestner, R., & Ryan, R.M. (2001). Extrinsic rewards and intrinsic motivation in education: Reconsidered once again. Review of Educational Research, 71, 1-27.
Dignath, C., & Buttner, G. (2008). Components of fostering self-regulated learning among students: A meta-analysis on interventions studies at primary and secondary school level. Metacognition & Learning, 3, 231-264.
Elley, W.B., & Reid, N. (1969). Progressive Achievement Tests: Reading comprehension. Wellington, New Zealand: New Zealand Council for Educational Research.
Goswami, U. (2008). Reading, complexity and the brain. Literacy, 42(2), 67-74. Guthrie, J.T., & Humenick, N.M. (2004). Motivating students to read: Evidence for classroom practices that increase reading motivation and achievement. In P.
McCardle & V. Chhabra (Eds.), The voice of evidence in reading research (pp. 329-354). Baltimore: Paul H. Brookes.
Guthrie, J.T., & Wigfield, A. (1999). How motivation fits into the science of reading. Scientific Studies of Reading, 3, 199-205.
Guthrie, J.T., & Wigfield, A. (2000). Engagement and motivation in reading. In M.L. Kamil, P.B. Mosenthal, P.D. Pearson, & R. Barr (Eds.), Handbook of reading research (Vol. 111, pp. 403-422). Mahwah, NJ: Erlbaum.
Hattie, J. (2009). Visible learning: A synthesis of over 800 meta-analyses relating to achievement. Abingdon, United Kingdom: Routledge.
Jacobs, J.E., & Paris, S.G. (1987). Children's metacognition about reading: Issues in definition, measurement and instruction. Educational Psychologist, 22, 225-278.
Jenkins, J.R., Fuchs, L.S., van den Broek, P., Espin, C., & Deno, S.L. (2003). Sources of individual differences in reading comprehension and reading fluency. Journal of Educational Psychology, 95(4), 719-729.
Juel, C. (1988). Learning to read and write: A longitudinal study of 54 children from first through fourth grades. Journal of Educational Psychology, 80(4), 437-447.
Logan, S., & Johnston, R. (2010). Investigating gender differences in reading. Educational Review, 62, 175-187.
Marsh, H.W. (1990). Influences of internal and external frames of reference on the formation of maths and English self-concepts. Journal of Educational Psychology, 82, 107-116.
Marx, H. (1998). Knuspels Leseaufgaben, Forms A & B. Gottingen, Germany: Hogrefe.
McLain, K.V.M., Gridley, B.E., & McIntosh, D. (1991). Value of a scale used to measure metacognitive reading processes. Journal of Educational Research, 85, 81-87.
Miller, S.D., & Faircloth, B.S. (2009). Motivation and reading comprehension. In S.E. Israel & G.G. Duffy (Eds.), Handbook of research on reading comprehension (pp. 307-322).
Ming Chui, M., & McBride-Chang, C. (2006). Gender, context and reading: A comparison of students in 43 countries. Scientific Studies of Reading, 10(4), 331-362.
Moller, J., & Koller, O. (2004). Die Genese akademischer Selbstkonzepte: Effekte dimensionaler und sozialer Vergleiche [On the development of academic self-concepts: The impact of social and dimensional comparisons.] Psychologische Rundschau, 55, 19-27.
Mossenson, L., Hill, P., & Masters, G. (1995). Tests of Reading Comprehension (TORCH). Melbourne, VIC: Australian Council for Educational Research.
National Institute of Child Health and Human Development. (2000). Report of the National Reading Panel. Teaching children to read: An evidence-based assessment of the scientific research literature on reading and its implications for reading instruction (NIH Publication No. 00-4769). Washington, DC: U.S. Government Printing Office.
Oakhill, J., Cain, K., & Bryant, P.E (2003). The dissociation of word reading and text comprehension: Evidence from component skills. Language and Cognitive Processes, 18, 443-468.
Paris, S.G., Lipson, M.Y., & Wixson, K.K. (1994). Becoming a strategic reader. In R.B. Ruddell, M.R. Ruddell, & H. Singer (Eds.), Theoretical models and processes of reading (pp. 788-811). Newark, DE: International Reading Association.
Paris, S.G., & Paris, A.H. (2001). Classroom applications of research on self-regulated learning. Educational Psychologist, 36(2), 89-101.
Pekrun, R. & Zirngibl, A.C. (2004). Schulermerkmale in Mathematik. [Student characteristics in mathematics]. In PISA-Konsortium Deutschland (Hrsg.): PISA 2003. Der Bildungsstand der Jugendlichen in Deutschland--Ergebnisse des zweiten internationalen Vergleichs (Seiten 191-210) [PISA 2003. The level of education of young people in Germany--Results of the second international comparison (pp. 191-210). Munster: Waxmann.
Pintrich, P.R., & Schrauben, B. (1992). Students' motivational beliefs and their cognitive engagement in academic tasks. In D. Schunk & J. Meece (Eds.), Student-perceptions: Causes and consequences (pp. 149-183). Hillsdale, NJ: Erlbaum.
Pressley, M. (2002). Comprehension strategies instruction: A turn-of-the-century status report. In C.C. Block, & M. Pressley (Eds.), Comprehension instruction: Research based best practices (pp. 11-27). New York: Guilford.
Pressley, M. (2006). Reading instruction that works (3rd ed.). New York: Guilford.
Pressley, M., & Harris, K.H. (2006). Cognitive strategy instruction: From basic research to classroom instruction. In P.A. Alexander, & P.H. Winne (Eds.), Handbook of educational psychology (2nd ed., pp. 265-286). Mahwah, NJ: Erlbaum.
Puustinen, M., & Pulkkinen, L. (2001). Models of self-regulated learning: A review. Scandinavian Journal of Educational Research, 45(3), 269-286.
Renninger, K.A., Hidi, S., & Krapp, A. (1992). The role of interest in learning and development. Hillsdale, NJ: Erlbaum.
Rider, N., & Colmar, S. (2006). Reading achievement and reading self-concept in Year 3 students. In P.L. Jeffery (Ed.), AARE Education Research. Creative Dissent: Constructive Solutions (pp. 1-Paper No). Parramatta, 27 Nov-1 Dec 2005.
Roeschl-Heils, A., Schneider, W., & van Kraayenoord, C.E. (2003). Reading metacognition and motivation: A follow-up study of German students in Grades 7 and 8. European Journal of Psychology of Education, 18(1), 75-86.
Samuels, S.J. (2006). Reading fluency: Its past, present, and future. In T. Rasinski, C. Blachowicz, & K. Lems (Eds.), Fluency instruction: Research-based best practices (pp. 7-20). New York: Guilford.
Schiefele, U. (1991). Interest, learning and motivation. Educational Psychologist, 26, 299-323.
Schlagmuller, M., Vise, M., & Schneider, W. (2001). Zur Erfassung des Gedachtniswissens bei Grundschulkindern: Konstruktionsprinzipien und empirische Bewahrung der Wurzburger Testbatterie zum deklarativen Metagedachtnis. [The acquisition of primary school children's knowledge about memory: Design principles and empirical verification of the Wurzburg Test Battery of Declarative Metamemory]. Zeitschrift fur Entwicklungspsychologie und Padagogische Psychologie, 33(1), 91-102.
Schneider, W., & Pressley, M. (1997). Memory development between two and twenty (2nd ed.). Mahwah, NJ: Erlbaum.
Snow, C.E., Burns, M.S., & Griffin, P. (Eds.). (1998). Preventing reading difficulties in young children. Washington, DC: National Academy Press.
Sweet, A.P., & Snow, C.E. (2003). Reading for comprehension. In A.P. Snow C.E. Snow (Eds.), Rethinking reading comprehension (pp. 1-11). New York: Guilford Press.
Taboada, A., Tonks, S.M., Wigfield, A., & Guthrie, J.T. (2009). Effects of motivational and cognitive variables on reading comprehension. Reading and Writing, 22, 85-106.
Taylor, B.M., Pearson, P.D., Garcia, G.E., Stahl, K., & Bauer, E. (2006). Improving students' reading comprehension. In K. Stahl (Ed.), Seeking understanding in how to teach reading: Selected works by Steven Stahl (pp. 303-315). New York: Guilford.
Torgesen, J.K. (2002). The prevention of reading difficulties. Journal of School Psychology, 40, 7-26.
Tunmer, W.E., & Chapman, J.W. (2002). The relation of beginning readers' reported word identification strategies to reading achievement, reading-related skills, and academic self-perceptions. Reading and Writing: In Interdisciplinary Journal, 15, 341-358.
van Kraayenoord, C.E. (1996). The Interest in Reading Scale. Unpublished scale, The University of Queensland, Fred and Eleanor Schonell Special Education Research Centre, Brisbane, Australia.
van Kraayenoord, C.E., & Schneider, W. (1999). Reading achievement, metacognition, reading self-concept and interest: A study of German students in Grades 3 and 4. European Journal of Psychology of Education, 14, 305-324.
Vellutino, F.R., Scanlon, D.M., & Tanzman, M.S. (1994). Component of reading ability: Issues and problems in operationalizing word identification, phonological coding, and orthographic coding. In G.R. Lyon (Ed.), Frames of reference for the assessment of learning disabilities: New views on measurement issues (pp. 279-332). Philadelphia, PA: Brookes Publishing Co.
Vise, M., Schlagmuller, M., & Schneider, W. (1998). Wurzburg Metamemory Test (rev. ed.). Wurzburg, Germany: Universitat Wurzburg, Institut fur Psychologie.
Wang, J.H.Y., & Guthrie, J.T. (2004). Modeling the effects of intrinsic motivation, extrinsic motivation, amount of reading, and past reading achievement on text comprehension between U.S. and Chinese students. Reading Research Quarterly, 39, 162-186.
Wigfield, A. (1997). Reading motivation: A domain-specific approach to motivation. Educational Psychologist, 32, 59-68.
Winne, P.H. (2005). A perspective on state-of-the-art research on self-regulated learning. Instructional Science, 33(5/6), 559-565.
Woodcock, R.W. (1987). Woodcock Reading Mastery Tests--Revised, Form H. Examiner's Manual. Circle Pines, MN: American Guidance Service.
Woolley, G.E. (2007). A comprehension intervention for children with reading comprehension difficulties. Australian Journal of Learning Disabilities, 12(1), 43-50.
Christina E. van Kraayenoord, (1) Andrea Beinicke, Matthias Schlagmuller & Wolfgang Schneider (2)
(1) The University of Queensland, (2) University of Wurzburg, Germany
Table 1. Mean ages (and standard deviations) of sample in years and months by gender Gender Grade 3 Grade 4 Males 9.14 (0.41) 10.08 (0.36) N = 30 N = 34 Females 9.16 (0.45) 10.07 (0.40) N = 31 N = 44 Table 2. Results of t-tests [M.sub.male] [SD.sub.male] [M.sub.female] READCOM -0.34 1.95 0.29 TEACHJUDGE -0.14 1.08 0.12 TORCH -0.20 1.05 0.17 WORDID -0.13 1.15 0.11 META -0.17 1.70 0.14 AWARE -0.03 1.00 0.02 METAMEM -0.14 1.03 0.12 SELFCON -0.23 1.05 0.19 INTEREST -0.03 1.11 0.02 [SD.sub.female] t p d READCOM 1.67 -2.06 .04 0.35 TEACHJUDGE 0.91 -1.57 .12 0.27 TORCH 0.92 -2.21 .03 0.38 WORDID 0.83 -1.35 .18 0.23 META 1.58 -1.10 .27 0.19 AWARE 1.00 -0.31 .76 0.05 METAMEM 0.96 -1.51 .13 0.26 SELFCON 0.91 -2.53 .01 0.43 INTEREST 0.89 -0.31 .76 0.05 Table 3. Correlations 1 2 3 4 1 READCOM 1 2 TEACHJUDGE 92 ** 1 3 TORCH 92 ** .69 ** 1 4 WORDID .76 ** .73 ** .66 ** 1 5 METAMEM .46 ** .46 ** .39 ** .30 ** 6 AWARE .34 ** .34 ** .28 ** 27 ** 7 SELFCON .48 ** .42 ** .47 ** .41 ** 8 INTEREST 0.12 0.08 0.14 .20 * 5 6 7 8 1 READCOM 2 TEACHJUDGE 3 TORCH 4 WORDID 5 METAMEM 1 6 AWARE .35 ** 1 7 SELFCON .21 * .19 * 1 8 INTEREST -0.03 .19 * .36 ** 1 * p < .05; ** p < .01 Note. These correlations refer to the means of the z-scores. Table 4. Stepwise multiple regression Variable B SE B P [sr.sup.2] (incremental) Step 1 Word identification 0.70 0.05 .76 *** .58 *** Step 2 Word identification 0.62 0.05 .67 *** .06 *** Metacognitive knowledge 0.14 0.03 .26 *** Step 3 Word identification 0.56 0.05 .61 *** Metacognitive knowledge 0.13 0.03 .24 *** .02 *** Reading self-concept 0.16 0.05 .17 *** [R.sup.2] = .66 ***, R = .81 ***. ** p < .01, *** p < .001. Note. R = Multiple correlation [R.sup.2] = Multiple correlation squared