Evolution of pauses during reading in schoolchildren with different reading comprehension trajectories

Introduction

From a phonetic perspective, prosody is understood as the set of suprasegmental features characteristic of speech: stress, intonation, tone, rhythm, melody, pauses, rate of speech, and voice quality (Llisterri, 2021). From a complementary perspective, Joannette et al. (2008) see it as a communicative resource that enables the understanding and expression of communicative intentions through a cognitive process based on suprasegmental features of intonation, variations in intensity and pauses. These traits are also evident in reading aloud and are good indicators of the level of development of fluent reading skills (Dowhower, 1991; Schreiber, 1991).

Reviews and meta-analyses report that prosodic performance in fluent readers corresponds to reading with greater amplitude and variation in pitch (F0), with grouping of words into meaningful units (phrases, sentences), and with pauses whose location and duration do not interfere with the meaning of the sentences (Godde et al., 2020; Wolters, et al., 2022). Research shows that, after the initial learning of reading, poor prosodic performance is linked to low reading comprehension, even in children with decoding skills appropriate for their age (Groen et al., 2019; Miller & Schwanenflugel, 2008; see the review by Roldán et al., 2022, based on studies with Spanish-speaking children). Furthermore, prosodic performance in oral reading can predict which students belong to the group with higher or lower reading comprehension skills (Groen et al., 2019).

Pauses are one of the prosodic aspects of oral reading that first differentiates between schoolchildren with good and poor reading comprehension (Álvarez-Cañizo et al., 2015; 2018). Nevertheless, there are still few longitudinal studies on this subject. In this context, this study analyses the evolution of the use of pauses in oral reading by schoolchildren in Chile with different levels of reading comprehension and establishes comparisons with adult readers as a reference group. To contextualise the study, the following section explains the different types of pauses. In addition, a brief review of the state of the art on the use of pauses in reading aloud and their relationship to reading comprehension is presented.

Reference framework

In spoken language, a pause corresponds to a ‘silence or vocalisation inserted into speech’ (Gil-Fernández, 2007, p. 544). Their location, frequency, and duration are associated with breath-voice coordination and articulatory planning (Godde et al., 2022), with the delimitation of the discourse units that structure speech, and with discourse planning (Hernández-Rodríguez and Soto-Barba, 2018). It is possible to distinguish between loud or full pauses and silent or empty pauses (Llisterri, 2021). Voiced sounds are vocalisations (‘mm’) or lengthening of vowels and consonants in words, while silent pauses correspond to the interruption of phonation.

When reading aloud, pauses are less frequent than in speech. Furthermore, silent pauses predominate over vocal pauses (Etxebarria et al., 2016; Gaminde et al., 2012; 2014). The research mainly focuses on the former, which can be grammatical or ungrammatical. Grammatical punctuation marks serve an important demarcative function, generally guided by punctuation. In other words, they respond more to syntactic prosody than to expressive prosody (Erekson, 2010), as they allow the reader to read while respecting the syntactic boundaries and meaning of the text. As a secondary or optional measure, the reader can also use pauses for expressive purposes, for example, by lengthening certain segments of words (‘you're sooo selfish’). On the other hand, agrammatical pauses are inadequate pauses between words, causing a pause in the sentence structure and potentially altering the meaning of what is read (Gil-Fernández, 2007). For less skilled readers, intrusive pauses can also be found, which can cause phenomena such as vowel or consonant lengthening, generally related to greater difficulty in reading words (‘astonnn...ished’) (Borzone-de-Manrique & Signorini, 2000; Schwanenflugel et al., 2004).

Some research shows that the ability to manage pauses well during oral reading develops relatively quickly in the early years of schooling, probably because it is one of the prosodic aspects that benefits most from good word recognition (Cuadro et al., 2021; Godde et al., 2020; Groen et al., 2019; Kim et al., 2021; Wolters et al., 2022). The study by Godde et al. (2022) is noteworthy, as it analyses the pattern of respiratory and syntactic-prosodic pauses in 295 French-speaking students between the second and seventh years of primary education, and compares it with that of 20 adults. The results revealed that second-year participants made more non-grammatical pauses, breathed more frequently, and produced more delays between inhalation and phonation. Fourth-year students, on the other hand, planned their breathing pauses better and were more guided by the score. For their part, seventh graders planned their pauses just like adults. The study concludes that proper pause management is acquired gradually between second and seventh grade, with punctuation playing an important role as a guide.

In Spain, Álvarez-Cañizo et al. (2018) describe pauses marked by punctuation marks, and pauses between and within words in 20 third-year primary school students, 20 fifth-year students and a group of adults. The total number of pauses and their duration were higher in the third-year group. Some characteristics of adult reading were found - mainly in terms of the length of pauses - among fifth-year students. Subsequently, in a longitudinal study with 40 students, Álvarez-Cañizo et al. (2020) found that inadequate pauses (not consistent with punctuation) decreased significantly between the first and second year, while their duration gradually decreased towards the third year. The number of adequate pauses (in line with the score) remained the same, but their duration decreased significantly between the second and third years.

For their part, Etxebarria et al. (2016) analysed the performance of 100 bilingual university students (Basque-Spanish) who read a news article aloud. Silent pauses were the most frequent and varied greatly in length. Although only 6% of the total corresponded to ungrammatical pauses, 35% of participants made some ungrammatical pause or some intrusive pause within a word.

Method

The design is descriptive and longitudinal. Data collection was carried out in two sessions in the 4^th and 6^th year, with no intervention between the two measurements.

Results

First, data on reading comprehension for each student profile is presented. Table 1 shows the descriptive statistics corresponding to the means (M) and standard deviations (SD) at the two moments of test administration.

Secondly, to answer the research question, the results of the analysis of pauses are presented: total, adequate, inadequate, and subtypes. The descriptive statistics (M and SD) referring to the total number of pauses in each group are shown in table 2.

The analysis revealed significant differences according to profile, both in 4^th year (H = 24.4, p < .01, ${\eta }^2$ = 0.7) and in 6^th year (H = 26.1, p < .01, ${\eta }^2$ = 0.7). In both years, the differences between the total number of pauses made by the group with good comprehension (profile 1) and those with low comprehension (profile 3) were statistically significant, as shown by the letters. The profile 2 group ranked in the middle, with no statistically significant differences from the other two groups.

As for university students, the differences were significant with the three school profiles, both in 4^th and 6^th year. Finally, the total number of pauses decreased significantly in 6^th year in all three profiles (profile 1: W = 2.1, p = .03, ${\eta }^2$ = 0.5; profile 2: W = 2.5, p = .01, ${\eta }^2$ = 0.6; profile 3: W = 2.7, p <.01, ${\eta }^2$ = 0.6).

The next step was to examine the adequate pauses. The corresponding descriptions are presented in table 3.

No significant intergroup differences were found in the number of adequate pauses in either 4^th year (H = 1.6, p = .66, ${\eta }^2$ = 0.0) or 6^th year of primary education (H = 0.2, p = .41, ${\eta }^2$ = 0.0). No differences were found when comparing 4^th and 6^th year in each profile. These pauses correspond almost entirely to silences coinciding with punctuation marks, with very few instances of lengthening as expressive pauses. The results for these subtypes of pauses are shown in table 4.

In contrast, significant differences were found in the total number of inadequate pauses in 4^th year (H = 25.3, p < .01, ${\eta }^2$ = 0.7) and 6^th year (H = 26.5, p < .01, ${\eta }^2$ = 0.8). The descriptions and results of the contrasts (letters) indicating where the differences are located are presented in table 5.

In 4^th year, there was a significant difference between the total number of inadequate pauses taken by the group belonging to profile 1 and profile 3, which is the group that takes the most pauses of this type. In 6^th year, this latter group continues to take the most inadequate pauses, but the difference with their peers in profile 1 was not significant. However, in this second measurement, all three groups significantly reduced the number of inadequate pauses made in 4^th year (profile 1: W = 2.3; p = .02, ${\eta }^2$ = 0.5; profile 2: W = 2.5, p = .01, ${\eta }^2$ = 0.6; profile 3: W = 2.7, p = .00, ${\eta }^2$ = 0.6). With regard to university students, who took few inadequate pauses, the three groups of schoolchildren showed significant differences in both the 4^th and 6^th year of primary education.

Table 6 shows the distribution of inadequate pauses in each subtype.

There were significant differences in the number of inadequate pauses of the silences type in 4^th year (H = 28.6, p < .01, ${\eta }^2$ = 0.8) and in 6^th year (H = 25.2, p < .01, ${\eta }^2$ = 0.7). In both year, the profile 3 group, with lower reading comprehension skills, made significantly more inadequate silences than the group with the best reading comprehension skills. The same is repeated in 6^th year. With regard to differences with university students, these were significant in the 4^th year for profiles 2 and 3, but not for profile 1. As for the decrease in inadequate silences in 6^th year, it was significant for profile 2. (W = 2.5, p = .01, ${\eta }^2$ = 0.6) and profile 3 (W = 2.7, p < .01, ${\eta }^2$ = 0.6), but not for the group with good comprehension, which already had few inadequate pauses in 4^th year.

As for inadequate pauses such as prolongations, intergroup differences were significant in 4^th year (H = 19, p < .01, ${\eta }^2$ = 0.5) y in 6^th year (H = 23.9, p < .01, ${\eta }^2$ = 0.7). In both years, profiles 1 and 3 differ from each other, with profile 3 incurring a significantly higher number of prolongations. In addition, the three groups made significantly more prolongations than the university students. Finally, in 6^th year, all three groups reduced these inadequate pauses, but only for profile 1 was the difference significant (W = 2.2, p = .03, ${\eta }^2$ = 0.5). As for profile 2, it was found that W = 1.6, p = .11, ${\eta }^2$ = 0.4. For profile 3, W = 0.9, p = .41, ${\eta }^2$ = 0.2.

Finally, the analysis revealed significant differences in pauses caused by hesitations in both 4^th year (H = 20.3, p < .01, ${\eta }^2$ = 0.6), and in 6^th year (H = 20.0, p < .01, ${\eta }^2$ = 0.6). The differences were significant between each profile and university students, but not between the profiles. The comparison between 4^th and 6^th year showed that the decrease in the number of hesitations was significant for schoolchildren with profile 2 (W = 2.4, p = .02, ${\eta }^2$ = 0.6), but it was not enough for schoolchildren in profile 1 (W = 2.0, p = .05, ${\eta }^2$ = 0.5) or profile 3 (W = 1.5, p = .14, ${\eta }^2$ = 0.4).

Discussion and conclusions

The aim of this study is to analyse, from a longitudinal perspective, pauses in oral reading in students with three different profiles of reading comprehension development: good comprehension in 6^th and 6^th year; low comprehension in 4^th year but good in 6^th year; and low comprehension in both years. We analysed their performance and compared it with a group of university students.

In all three profiles, there was a decrease in the total number of pauses in 6^th year, but the difference between the highest and lowest performing groups in reading comprehension remained significant. Similarly, the three groups showed significant differences compared to university students. In this regard, our results are consistent with previous studies such as that by Álvarez Cañizo et al. (2018), which found that third-year children took more pauses than fifth-year children, although both groups took more pauses than the adult participants.

We then quantified the adequate and inadequate pauses. The three profiles showed practically the same number of adequate pauses in both years, with a performance similar to that of university students and closely matching the number of punctuation marks in the text. This is consistent with research showing that the ability to plan pauses is acquired with significant support from scoring (Godde, 2022; Álvarez Cañizo et al., 2020). In this study, as in Godde’s (2022), we find that in 4^th year, children were already able to read while respecting the grammatical pauses marked by punctuation, even if their reading comprehension remained low. In other words, it seems that at this school level, most students are already aware of the role of punctuation marks and use them to pause where necessary. This explains why no differences or interactions were found between the year and profile variables, nor with university students, suggesting a certain ceiling effect that leaves no room for changes that could impact performance.

Inadequate pauses are a good indicator of reading skill development in school-age children and confirm that lower reading comprehension is linked to more ungrammatical or inadequate pauses in oral reading (Álvarez-Cañizo et al., 2015; Arcand et al., 2014; Borzone-de-Manrique & Signorini, 2000). This could be observed in 4^th year, as the frequency of inadequate pauses marks a significant gap between those who show better development in comprehension and those with lower reading comprehension, while profile 2 falls between the two groups. Subsequently, with two more years of schooling, we found positive developments in all three profiles. However, although all three groups demonstrated potential for improvement, the reduction in inadequate pauses was only significant for the profile with good comprehension in both years, which was the group that came closest to the performance of university students.

As for subtypes of inadequate pauses, silences were the most frequent. These intrusive pauses disrupt reading continuity and could reflect a lower efficiency in integrating words into the syntactic chain in the absence of punctuation or the need to plan the discourse articulation process. The group with the lowest reading comprehension scores made more of these types of pauses, although the difference with students in profile 2 was not statistically significant in either 4^th or 6^th year. The latter group reduced inadequate silences in 6^th year, but did not achieve the performance of adult readers.

Prolongations also make it possible to differentiate between the group with the best track record in comprehension and the group that showed its low performance. In 6^th year, all three groups reduced these pauses, although the differences were not significant compared to the first measurement. However, the decrease in prolongations in profile 1 allowed this group to catch up with the university students. As for hesitations, significant differences were only found among university students, and 4^th and 6^th year in profile 3.

In conclusion, the results support the initial prediction that differences in the development trajectories of reading comprehension may be linked to differences in the evolution of pause use in oral reading. Two factors are essential. On the one hand, the interpretation of punctuation marks, which allows children to make grammatical pauses without difficulty. On the other hand, the frequency of silences, hesitations and inadequate prolongations seems to be a relevant marker of the evolution of pause management. Taken together, these results suggest that the level of automaticity in reading continues to be a factor in the performance of schoolchildren with lower reading comprehension skills. Even though we selected participants who had no difficulty reading words and found no intergroup differences in this skill, its subsequent development could have been a factor linked to their own trajectories in reading comprehension and fluency. Phenomena such as prolongations and hesitations during reading aloud generate inadequate pauses that reflect a conscious effort to read words correctly, to the detriment of reading comprehension. In this sense, the improvement in understanding among students in profile 2 could be related to an improvement in this aspect.

In summary, this research, from an evolutionary perspective, helps to understand how the acquisition of pauses, a key element of reading fluency, varies in schoolchildren with different profiles of reading comprehension development. Added to this is the methodological relevance of the study, which incorporates a detailed analysis of pause subtypes and establishes a point of comparison with expert readers. Together, these elements contribute to a more accurate description of the characteristics that distinguish fluent reading in the intermediate stages of reading learning.

A practical implication of the study is that it suggests that oral reading should continue to be practised throughout primary education, as it provides important information about the individual development of reading skills in the classroom. Similarly, it is very important that children learn the meaning of punctuation marks in their early years at school and have opportunities to demonstrate how they are progressing in their use as a guide to making grammatical pauses when reading. However, achieving fluent reading, without intrusive or ungrammatical pauses, does not depend solely on that. Targeted practice and specific support are required for children to read without hesitation or prolongation, which seem to indicate that the automaticity of reading still needs to be improved in order to achieve good fluency.

One limitation of this research is that data for 6^th year correspond to 2021, when almost two years of remote learning had passed, during which many reading-related practices were interrupted or became less regular. Without this factor, the participants' potential for improvement would probably have been better reflected in the results.

As a projection, it is suggested to explore the link between word reading and pauses, taking into account their duration. We believe this variable is potentially relevant for exploring the cognitive effort involved in word decoding and articulatory planning. Other measures that could enrich future research include speech rate, intonation, syntactic grouping, and melodic variation. It would be valuable to explore how these prosodic parameters interact in reading learning, in order to advance towards a more comprehensive understanding of reading development in Chilean schoolchildren.