Exploring the Relationship between Reading Comprehension and Text Genres in Fifth and Sixth Grade ()
1. Text Genres
Texts are vehicles for knowledge acquisition. It is the reader who absorbs and transmits knowledge; thus, the relationship between the reader and the text is important and, for that reason, it is a subject that is taught at all levels of our educational system. The distinction of text genres—in narrative and informational—aims at their more detailed study and their better teaching.
The structure of narrative texts is more familiar to students, compared to the structure of informational texts, with the result being that they understand the former more easily than the latter (Clinton et al., 2020; McTavish, 2008; Pressley, 2000). In the US, research has shown that students are taught narrative texts four times more often than informational texts (Erickson, 1998); Duke (2004) adds that teachers in the first grades of elementary school dedicate only 3.6 minutes per day to the teaching of informational texts. The students’ limited contact with informational texts was considered as one of the reasons why the fourth and, mainly, fifth-graders present a declining course, or a slump, in the development and achievement of comprehension, as the texts gradually become more demanding (Sanacore & Palumbo, 2009; Best, Floyd, & McNamara, 2004).
Narrative and informational texts showcase different structural characteristics, which readers inadvertently or intentionally seek, as they rely on them in order to comprehend the texts, and also because they are indicators of the difficulties they possibly encounter in terms of reading comprehension (Collins et al., 2020; Best, Ozuru, Floyd, & McNamara, 2006; McNamara & Kintsch, 1996).
Narrative texts are characterized by the “grammar of the text”, i.e. a coherent textual structure, which is based on time sequence and cause-effect relationship (Graesser, Golding, & Long, 1991; de Beaugrande & Dressler, 1981).
The narrative structure, or textual scheme, or “grammar of the text” was categorized (Page & Stewart, 1985) into two models: a) the model based on the creation of logical inferences (Graesser, McNamara, & Louwerse, 2003), where the information required to understand the text may not be contained in it; b) the model based on the textual structure (Halliday & Hasan, 1976), according to which the reader, in order to understand a text, should create cognitive representations based only on information retrieved from the text (Kintsch, 2005; Graesser, McNamara, & Louwerse, 2003; Kintsch, 1988). Students with reading difficulties may not be able to efficiently utilize one of the two models, or both models at the same time.
Unlike narrative texts that are oriented to the action of the protagonist(s), informational texts are timeless, oriented to a logically structured subject. While informational texts present a variety of structures with which students may not be familiar (Graesser, McNamara, & Kulikowich, 2011), narrative texts follow a specific textual structure, with which they are usually acquainted from a very young age; Englert & Hiebert (1984) and Meyer & Freedle (1984) distinguish informational texts based on their structure in: 1) cause-effect, 2) descriptive, 3) logical sequencing, 4) comparison-contrast, and 5) problem-solving.
Another difference between informational and narrative texts is that the former offer readers new information, which they must understand in order to be able to transmit it, as this is the primary purpose of these texts (Pyle et al., 2017; Best, Rowe, Ozuru, & McNamara, 2005; Duke, 2004; Hall, 2004).
2. The Role of Text Genres in Reading Comprehension
At a very early age, children come into contact with reading through fairy tales or stories that they hear from their family environment. Thus, these simple in content texts, read repeatedly to children, constitute the means through which the process of comprehension is inadvertently being “learned” and further supported by the rich accompanying illustration (Araïjo & Costa, 2015). When children enter school, they move on to reading for the acquisition of knowledge, and to more demanding levels of comprehension, while text starts to be viewed as an educational tool (McNamara, Floyd, Best, & Louwerse, 2004). Gradually, students come into contact with different types of texts: texts in school textbooks, texts from books or magazines, electronic texts, narrative, descriptive, informative, argumentative, advertising texts, etc.
Fox & Alexander (2009: pp. 230-232), in a review of several studies, argue that readibility, the text genre, the textual structure, coherence, the communication channel (school textbooks, electronic information) and printing are the most important textual factors affecting reading comprehension. There are several researchers (Loyd & Steele, 1986; Singer & Ruddell, 1970) who have pointed out the significance of the underlying message of the text, giving the reader the role of the “decoder” and “recreator” of the text’s meaning.
According to constructivist theories (Kintsch, 1998, 2013), students primarily understand the “surface l code”, i.e. the words of the text; then, they create the “text base”, i.e. organize the “ideas”—a process which is influenced by the syntactic structure of the text and the use of cognitive strategies; finally, they form a “situational model” by utilizing their prior knowledge and various metacognitive strategies (Perfetti, Landi, & Oakhill, 2005).
The significance of textual factors, along with the difficulties in reading comprehension that are mainly posed for fourth-grade students [fifth-grade based on Greek curricula] by scientific and informational texts, were studied by Sweet & Snow (2003), as well as Chall & Jacobs (2003).
Therefore, it becomes clear that familiarity with the structure of the text genre, along with the organization of the text, constitutes not only a variable for assessment, but also a teaching aim that is necessary for successful reading comprehension.
3. Comprehension Levels
The ways in which a text can potentially be interpreted are infinite, but they decrease dramatically when we focus on what someone has understood from that text. In recent years, we have seen multiple research that deals with the process of reading comprehension, as well as the resulting outcome, which can be assessed in various ways (Alderson, 2000). After all, the “potential meaning” of the text is the result of the reader’s interaction with it.
King (2007) distinguishes two general levels of comprehension: literal/shallow comprehension, which concerns the information (topic sentences, details) that a person can recall and which are inherent in the text; and the inferential/profound comprehension, which requires the reader’s critical thinking and deeper involvement with the text, as it is determined by high coherence, rich integration and plausible cognitive representations.
The reader can make simple sentential and intersentential conjunctions almost unconsciously, as well as “intratextual inferences” (Rayner & Reichle, 2010); these constitute the first category of inferences. The second category entails more complex inductions (causal, predictive, processing), which require the connection of the new information with the reader’s pre-existing knowledge and lead to the central meaning of the text (Dickens & Meisinger, 2017; van den Broek et al., 2005; Kintsch, 1998; Graesser, Singer, & Trabasso, 1994); these are the “intertextual inferences” (Krokou, 2011; Chiesi, Spilich, & Voss, 1979).
Studies analysing factors that influence reading cite vocabulary knowledge as one of the most important variables (Ouellette & Beers, 2010). Alderson (2000), invoking Johnston’s (1983) study, takes the view that vocabulary knowledge relates more to the reader’s background knowledge than to a “vocabulary” knowledge. Also, Wright & Cervetti (2017) point out that the comprehension of texts is enhanced by the active teaching of new words [textual approach] rather than by a passive juxtaposition of their conceptual clarification.
In this study we will refer to literal comprehension, vocabulary comprehension, intratextual comprehension and intertextual comprehension, as they are precisely defined, absolutely linked to strategies that children have learned or can be taught with appropriate intervention programs, and have been utilized successfully in various modern standardized tests (Critical Reading Inventory-2: CRI-2: Applegate, Quinn, & Applegate, 2008; Gray Oral Reading Test: Gort-4: Wiederholt & Bryant, 2001).
4. Purpose of the Research
This research aims to investigate whether the reading comprehension of students in the fifth and sixth grade is influenced by the text genre to which the text belongs, and in particular, whether there are differences in the comprehension of narrative and informational texts between the two grades. In addition, it explores whether students’ reading comprehension is differentiated by text genre in terms of the four (4) distinct levels: 1) vocabulary comprehension; 2) literal comprehension; 3) intratextual comprehension; and 4) intertextual comprehension. In Greece, there has been no prior empirically supported research of such kind that focuses on the need to clarify the interaction between the different comprehension levels, as well as on the ways with which they affect the process of comprehension and augment global knowledge.
5. Method
This research is part of a broader research aimed at assessing the reading comprehension of students with and without learning disabilities. A test was developed during initial research stages, in order to detect the level of reading comprehension (Krokou, 2011, 2018), and it was also used in this research.
Two questions are examined; the first is about the effect of text genres on reading comprehension, and the second concerns the impact of reading comprehension indicators, such as vocabulary, textbase indicator, inferential/intratextual indicator and intertextual indicator. Furthermore, the correlation between the four indicators of reading comprehension is considered. Relevant data deriving from the students of both grades were analyzed with the use of t-test and ANOVA.
6. Participants
The survey involved 1614 students from the fifth and sixth grade of elementary schools in Attica. Of the 1614 subjects in the sample, 824 (51.1%) were boys, and 790 (48.9%) were girls. The ages of the participants range from 9 years and 10 months to 14 years and 5 months, with an average of 11 years and a standard deviation of about 8 months (7.75).
Participating students were selected from the fifth (830 students, 51.4%) and the sixth grade (784 students, 48.6%) of elementary schools in the prefecture of Attica. It should be noted that all students of the selected classes participated in the study, including foreign students who had a good knowledge of the Greek language, and students with low performance due to the existence of neurobiological or genetic disorders.
7. The Reading Comprehension Assessment Test
The test for the assessment of reading comprehension consists of two texts, one narrative and one informational, of an equivalent level (Krokou, 2011).
In particular, the narrative text consists of 290 words and the informational text of 326 words. The number of sentences is similar in both texts (Krokou, 2011). The size of the texts is also consistent with the texts that children come into contact with in the classroom (Keenan, Betjemann, & Olson, 2008; Ozuru et al., 2007; Alderson, 2000). Still, both texts are specific, the narrative text refers to actual historical event, and the informational text to an existing animal, and they have been selected in a manner consistent with the interests of children of this age (Alderson, 2000: p. 62).
Both texts contain no unfamiliar words for fifth- and sixth-graders. In the narrative text there are few specimens of figurative speech (similes’, metaphors, entrenched expressions), the meaning of which can easily be concluded from the context. The informational text does not contain specimens of figurative speech, and the unknown words that the children may have are either clearly explained in an adjacent sentence, or their meaning is revealed in the context.
After the end of the texts, 50 comprehension questions follow, which have been designed on the basis of the four levels of comprehension: vocabulary comprehension, literal comprehension, intratextual comprehension, and intertextual comprehension (Krokou, 2011, 2018).
8. Process
As regards the administration of the reading comprehension assessment test, students first read the text and then are asked to answer the comprehension questions. In particular, the test is designed so that children read the text first and then answer the multiple-choice questions that follow it. Children are given the opportunity to refer to the text in order to clarify any question they wish (Sadoski & Paivio, 2007). This approach of comprehension assessment was selected because our interest does not lie in exploring what the students remember or how much knowledge they gain (Ozuru et al., 2007; Alderson, 2000), but in examining the availability or elimination inconsistencies of cognitive representations created during the first reading (Stine-Morrow et al., 2004; Alderson, 2000).
The duration of the test was approximately forty-five minutes. Each correct answer was scored with one (1) point, and each wrong answer with zero (0) points.
9. Findings
Initially, the difficulty indicator of the questions was studied, which is the percentage of correct answers for each question. The closer the indicator gets to 100%, the easier a question is (a question with an indicator of 100% has been answered by everyone), and the closer the indicator approaches 0%, the more difficult a question is (a question with an indicator of 0% has not been answered by anyone). If the indicator is around 50% (medium difficulty), the question is considered better and more effective. Many researchers consider that difficulty indicators should range from 20% to 80% (Kline, 2000). However, earlier studies suggest indicator values between 30% or 40% and 70% (Allen & Yen, 1979; Anastasi, 1982), as well as between 10% and 90% (Walsh & Benz, 1990). Almost everyone, however, agrees that the average indicator of difficulty should lie between 50% and 60%.
The statistical analyses show that the mean of the difficulty indicator (item difficulty) for the overall test is 68.20%; in particular, 65.80% for the fifth grade and 71.60% for the sixth grade.
9.1. Narrative and Informational Genres
In order to answer the first question about the effect of text genres on reading comprehension, relevant data deriving from the students of both classes were analyzed with the use of t-test and ANOVA. The results are presented in Table 1 and Table 2.
9.2. Reading Comprehension Indicators
In order to answer the second question concerning the impact of reading comprehension indicators, a statistical analysis of t-test independent samples was performed; the results are presented in Table 3. This analysis shows that the
![]()
Table 1. T-test analysis results for the narrative and the informational genre.
![]()
Table 2. ANOVA results for the narrative and the informational genre.
![]()
Table 3. T-test analysis results for reading comprehension indicators.
average of the vocabulary indicator values of sixth-graders (M = 0.767, SD = 0.22) is higher than those of fifth-graders (M = 0.685, SD = 0.22), but this difference is not statistically significant. Correspondingly, the textbase indicator of sixth-graders (M = 0.888, SD = 0.20) is higher than that of fifth-graders (M = 0.818, SD = 0.22), and this difference is statistically significant for p ≤ 0.05. The inferential indicator of sixth-graders (M = 0.717, SD = 0.17) is higher than that of fifth-graders (M = 0.672, SD = 0.18), and this difference is marginally statistically significant (p = 0.048). The intertextual indicator of sixth-graders (M = 0.624, SD = 0.16), is higher than that of fifth-graders (M = 0.566, SD = 0.16), and this difference is statistically significant for p ≤ 0.05.
ANOVA was performed later on, the results of which show that the two groups of students present statistically significant differences in terms of reading indicators per school grade (Table 4). This means that the effect of the students’ grade of attendance is statistically significant for the vocabulary indicator (F = 54.344, p = 0.000), for the textbase indicator (F = 45.271, p = 0.000), for the inferential/intratextual indicator (F = 28.952, p = 0.000), and for the intertextual indicator (F = 55.378, p = 0.000).
Finally, a correlation analysis (Table 5) was carried out, which shows that
![]()
Table 4. ANOVA results for reading comprehension indicators
![]()
Table 5. Correlations between the four reading comprehension indicators.
**Correlation is significant at the 0.01 level (2-tailed).
![]()
Figure 1. Reading comprehension indicators means chart.
there is a linear correlation between the four indicators of reading comprehension. Specifically, vocabulary indicator shows a satisfactory positive correlation with textbase indicator (r = +501), and moderate correlations with inferential indicator (r = 0.435) and intertextual indicator (r = 0.454). Also, textbase indicator shows strong correlations with inferential/intratextual indicator (r = 0.67) and intertextual indicator (r = 0.640). Finally, the inferential indicator also shows a strong correlation with the intertextual indicator (r = 0.636). All of the above correlations are statistically significant (sig = 0.000).
10. Discussion
The results of the statistical processing showed that textual genres affect students’ reading comprehension, as there was a statistically significant differentiation in students’ performance in both grades (Figure 1). In particular, all students presented a slightly higher performance regarding the informational text than the narrative. In addition, the students of the sixth grade achieved a higher overall performance in both texts than the students of the fifth grade. This finding, besides being plausible, is confirmed by corresponding studies that point out that, as students mature, they become more capable readers (Dickens & Meisinger, 2017).
In particular, the narrative text is not supported by the students’ prior knowledge, while the subject of the informational text is indeed supported by it. Although the discrepancy in comprehension between the two texts is considered statistically significant for both grades, it appears that students performed better in the comprehension of the informational text. Therefore, it is believed that the importance of prior knowledge should be highlighted, as it represents a research issue that concerns several scientists. Clinton et al. (2020) emphasize that no other factor affects comprehension as much as the individual’s prior knowledge of a subject, while Kintsch (2013) points out that when information is well organized in cognitive schemata and linked to the student’s prior knowledge, the situational model of the text becomes coherent, while comprehension is maximized.
In addition, research by Dickens & Meisinger (2017) confirms our own findings, as it highlights the effect of the general knowledge available to an individual on the comprehension of mainly informational texts.
Although both the students’ familiarity with the structure of text genres (textbase indicator and inferential indicator), and the contribution of knowledge they may have on a subject (intertextual indicator) were examined as for their impact on reading comprehension, the comparative results show that the most important factor for seamless reading comprehension is the background knowledge that is related to the text [strong correlation of inferential indicator with intertextual indicator (r = 0.636)]. These data are also supported by research conducted by McNamara & Kintsch (1996) on students of the same age as our sample; they emphasize the great importance of prior knowledge in reading comprehension, while, even in earlier research, such as this by Chiesi, Spilich, & Voss (1979), it is noted that prior knowledge often complements or enhances reading ability. However, an interpretation of the findings of the meta-analysis undertaken by Bogaerds-Hazenberg, Evers-Vermeul, & van den Bergh (2021) suggests that, although there are no significant data on their correlation, the textual structure should be systematically taught in school, considering that this is the best way to support the difficult transition to reading for the purpose of learning. This view is supported by research conducted by Graesser, McNamara, & Kulikowich (2011).
The effect of the textual structure/textbase indicator on reading comprehension was studied by McTavish (2008) and Pressley (2000), who point out that narrative texts are more comprehensible because of their familiar textual structure, unlike informational texts, the structure of which is not causal but linear, and does not favor the formation of comprehension-enhancing inferences (Best et al., 2006). The above finding may not have been fully verified by our research, as prior knowledge on the subject of the text seemed to affect reading comprehension more, but the performance of our students in comprehending the narrative text was equally important, leading us to conclude that the teaching of textual structure is necessary for proficient reading comprehension.
In addition, students of the sixth grade achieved higher performance in all reading comprehension indicators than students of the fifth grade, except for the vocabulary indicator. Regarding that indicator, the pupils of the sixth grade again performed better than the pupils of the fifth grade, but this difference is not considered statistically significant. The lowest performance of all students as for the vocabulary level concerned the comprehension of words whose meaning was derived from the utilization of prior academic knowledge. Research points out that vocabulary that is not familiar to students (Pyle et al., 2017; Best, Floyd, & Mcnamara, 2008; Hall, 2004), or vocabulary that has not been taught just before reading, or is not activated (Wright & Cervetti, 2017), leads to incomplete comprehension, as it lacks meaning on the reader’s end (Ouellette & Beers, 2010). This finding is consistent with the conclusions of other studies, which demonstrate the incomplete or limited teaching of vocabulary at school (Wright & Cervetti, 2017) as a predictive negative indicator for reading comprehension. The high correlation of the vocabulary indicator with the textual basis and less with prior knowledge via inferences, demonstrates that, at that time, students seek meaning within context and do not use their own background knowledge—either by utilizing their memory or by inferencing (Perfetti, Landi, & Oakhill, 2005: p. 240).
Meta-analytical research by Clinton et al. (2020), despite highlighting that most research data show that intertextual comprehension is higher in narrative texts than in informational texts due to a more coherent situational model, nevertheless concludes that the readability of the text or the familiar subject, in a significant number of studies, affects the successful creation of inferences for both text genres. This is confirmed by the high correlation of indicators in our research.
As regards literal comprehension, students performed well in both texts, with slightly better scores in the informational text. However, Clinton et al. (2020) report that—in narrative texts—students create inferences more easily on the textbase. The difference in findings may be due to the familiar subject of the informational text, which may contribute to the ability of students to more easily make connections between bits of information presented in the text. However, the researchers above suggest a correlation of readability with other textual characteristics, such as, but not limited to, vocabulary, textual indicators, or the use of pronouns, since some studies present relevant findings.
The high correlation of intratextual comprehension indicators with those of intertextual comprehension states that, when difficulties arise in the formation of intertextual inferences, i.e. in the utilization of students’ prior knowledge, there is limited use of cognitive and metacognitive strategies, to bridge the text’s conceptual gaps (Kintsch, 2013). Research conducted by Graesser, McNamara, & Kulikowich (2011) reports that intratextual comprehension is more demanding in informational texts, as students do not know or have not been taught how to manage their demanding syntactic structure, a finding confirmed by the strong correlation of the textbase indicator with the inferential indicator in this research.
In conclusion, many scientific opinions converge on the significance of prior knowledge in the development of reading comprehension (Clinton et al., 2020; McNamara & Kintsch, 1996), as it promotes the formation of higher levels of comprehension and facilitates the appropriation of new knowledge. However, the importance of the textual structure, which should continue to have a primary role in text teaching (Bogaerds-Hazenberg, Evers-Vermeul, & van den Bergh, 2021), should not be overstressed, as comprehension is promoted when students know how to approach a text based on the genre to which it belongs. What is an important predictor of successful comprehension is the child’s constant contact with different genres of printed speech from a very early age.
Research Limitations
The participants of the survey came only from the prefecture of Attica, so the results have been obtained from students attending schools of the largest urban center. In order to improve the representativeness of the results, research could be extended to schools in the wider Greek region, in semi-urban and rural areas.
In addition, a process could be created in a future study that would be able to measure more accurately the readability of texts.
Future Suggestions
The results of this study could be utilized by researchers and educators in order to create intervention programs for the teaching of structure in narrative and informational texts, with the use of cognitive and metacognitive strategies; these strategies could derive from the understanding of interaction mechanisms between the different levels of reading comprehension, and they could be applied to the identification and recollection of the most important data.
Moreover, the results could also be put into good use for the creation of a methodological tool that would generate strategies and techniques for logical inferencing, something very important for intra- and intertextual comprehension, especially in regard to informational texts.