Longitudinal associations between third‐grade teaching styles and sixth‐grade reading skills : a 3‐year follow‐up study

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY-NC 4.0 https://creativecommons.org/licenses/by-nc/4.0/ Longitudinal associations between third‐grade teaching styles and sixth‐grade reading skills : a 3‐year follow‐up study © 2022 the Authors Published version Tang, Xin; Kikas, Eve; Pakarinen, Eija; Laursen, Brett; Lerkkanen, Marja‐Kristiina Tang, X., Kikas, E., Pakarinen, E., Laursen, B., & Lerkkanen, M. (2022). Longitudinal associations between third‐grade teaching styles and sixth‐grade reading skills : a 3‐year follow‐up study. Journal of Research in Reading, 45(1), 157-169. https://doi.org/10.1111/1467-9817.12385 2022 Longitudinal associations between third-grade teaching styles and sixth-grade reading skills: a 3-year follow-up study Xin Tang School of Psychology, Central China Normal University, Wuhan, China and Faculty of Educational Sciences, University of Helsinki, Helsinki, Finland Eve Kikas School of Natural Sciences and Health, Tallinn University, Tallinn, Estonia Eija Pakarinen Department of Teacher Education, University of Jyväskylä, Jyväskylä, Finland Brett Laursen Department of Psychology, Florida Atlantic University, Boca Raton, FL, USA Marja-Kristiina Lerkkanen Department of Teacher Education, University of Jyväskylä, Jyväskylä, Finland Background: Most previous studies of teaching styles and reading skills have been cross-sectional. Longitudinal research is needed to clarify the direction of effects. The present longitudinal study examined the degree to which differences in teaching styles in the third grade predict the sixth-grade reading performance. The consistency of the findings was addressed by comparing results across students in two countries (Finland and Estonia). Methods: A total of 1,057 students (50.9% boys) were followed from the third to sixth grade. Teaching styles of third-grade teachers (N= 70) were examined as predictors of the development of reading (i.e., third-grade to sixth-grade reading fluency and comprehension). Results: Five patterns of third-grade teaching practices were found across two countries: child-centred style,teacher-directed style,child-dominated style,extreme child-centred style, and mixed child-centred and teacher-directed style (mixed teaching style). The mixed teaching style and the child-centred style in the third grade were related to the greatest increases in reading fluency from the third to sixth grade, over and above the contribution of age, gender and maternal education. Conclusions: The findings underscore the importance of the flexible use of child-centred and teacher-directed practices, which are both linked to the development of reading fluency during late primary school years. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. © 2022 The Authors. Journal of Research in Reading published by John Wiley & Sons Ltd on behalf of United Kingdom Literacy Association. Journal of Research in Reading, ISSN 0141-0423 DOI:10.1111/1467-9817.12385 Volume 00, Issue 00, 2022, pp 1–13 Keywords: teaching styles, reading development, early teaching effects, elementary students, reading comprehension Highlights What is already known about this topic •A person-oriented approach is an effective method to identify groups of teaching practices, known as teaching styles. •Teaching styles have been differentially linked to reading skills, but the associations have been limited to cross-sectional data. •Finnish teachers apply effective teaching styles more often than Estonian teachers. What this paper adds •This paper adds to previous studies by examining longitudinal associations between teaching styles and reading skills. •The study established the consistency of teaching styles between Finland and Estonia. •The study found that mixed teaching style is beneficial for reading fluency development. Implications for theory, policy or practice •The combinations of child-centred practices and teacher-directed practices are the most effective teaching styles for reading development. •The beneficial effect is consistent across Finland and Estonia. Early teaching practices exert a powerful influence over children’s later academic achievement (Hamre & Pianta, 2005; Stipek et al., 1998). Both classroom management practices and positive teacher–child relationships in early primary school have been tied to long-term academic success (Hamre & Pianta, 2001; Stipek & Byler, 2004). Most empirical studies of teaching adopt a variable-oriented approach, exploring the independent contribution of different teaching practices to child outcomes. Few studies have applied person-oriented methods to determine teacher behaviours that tend to cluster together to form teaching styles (e.g., Hu, Fan, LoCasale-Crouch, Chen & Yang, 2016; Kikas, Silinskas, Jõgi & Soodla, 2016). The present investigation extends findings from a previous person-oriented study (Tang, Kikas, et al., 2017) that linked specific constellations of teaching practices (i.e., teaching styles) to concurrent reading abilities. In the present study, we examined the role of early teaching styles in reading development from the third to sixth grade. We also tested our hypotheses in two countries (Finland and Estonia) to examine the effects in a cross-national context. TANG, KIKAS, PAKARINEN, LAURSEN & LERKKANEN © 2022 UKLA. Teaching practices, teaching styles and reading performance Much is known about the role of teaching practices in reading skills development (e.g., Connor et al., 2014; Pressley et al., 2001). Teaching practices are the behaviours that teachers employ in their daily interactions with children. Three broad perspectives define teaching practices. Child-centred practices, rooted in constructivist theories of learning and teaching, value children’s interests, provide individual supports and demonstrate warmth and affection (e.g., Vygotsky, 1978; for an overview, see Driscoll, 2005; Pressley et al., 2003). Teacher-directed (i.e., didactic) practices, based on behaviourism, emphasise basic skills acquisition and teachers’authority in classroom activities (Driscoll, 2005; Pressley et al., 2003). Child-dominated practices, with roots in maturation theory and a more extreme version of constructivism, overstress children’s autonomy and de-emphasise direct instruction (e.g., free play and discovery-based learning; Stipek & Byler, 2004). Although these conceptual distinctions have improved our understanding of the link between teaching and reading skills (Lerkkanen et al., 2016; Stipek et al., 1998), teachers in real classrooms do not strictly adhere to these prescriptions (Good, Wiley & Florez, 2009; Pressley et al., 2003). In authentic situations, teachers may utilise teaching practices in a more nuanced manner. In a previous publication (Tang, Kikas, et al., 2017), we documented variability in teaching practices. With the help of person-oriented analyses (Laursen & Hoff, 2006), we identified five different teaching styles (defined as the manifestations of theory-defined teaching practices) in third-grade classrooms: child-centred style, teacher-directed style,child-dominated style,mixed teaching style and extreme child-centred style. The first three teaching styles correspond with three theory-defined teaching practices introduced earlier, but the latter two represent practices that illustrate the diversity of real classroom experiences. Mixed teaching style occurred when child-centred practices and teacher-directed practices were equally presented. Extreme child-centred style represented an extremely high level of child-centred practices. The results further indicated that the child-centred style and the mixed teaching style were most effective in terms of reading performance, because they provide direct/explicit reading instruction while supporting children’s autonomy and interests. Teacher-directed style contributed to reading fluency as it offered direct instruction. Extreme child-centred style was associated with reading comprehension but not with reading fluency due to its support for the child’s interests. Child-dominated style was the least effective, which is perhaps not surprising given that minimal instruction and guidance were provided. This previous study was limited because it only examined cross-sectional associations between teaching styles and reading performance (Tang, Kikas, et al., 2017). Longitudinal studies are needed to shed light on the predictive role of teaching styles in the development of reading skills. In one longitudinal study (Tang, Pakarinen, Lerkkanen, Muotka & Nurmi, 2019), we tried to fill this gap by examining the role of first-grade teaching styles in the development of reading skills from the first to third grade. The results indicated that the mixed teaching style is most beneficial for reading skill development. Nevertheless, conclusions were limited by a focus on students and teachers in early primary school years from a single setting. The role of teaching styles on the reading development during late primary school years is still largely unknown. The late primary school years are a period of consolidation; achievement during these years is important for the transition to secondary school (Virtanen, Vasalampi, Kiuru, Lerkkanen & Poikkeus, 2020). These years are also noteworthy for the instructional changes that occur (see the succeeding text). TEACHING STYLE AND FOLLOW-UP READING © 2022 UKLA. In sum, the present study was designed to further our understanding of the long-term impact of teaching styles on reading development during late primary school years. Building on our previous research (Tang, Kikas, et al., 2017), we examined the role of third-grade teaching styles on reading development from the third to sixth grade. Moreover, we extended previous findings by examining the consistency of teaching styles and their outcomes in two countries: Finland and Estonia. Literacy instruction in Finland and Estonia during primary school years Both Finland and Estonia show high performance in international comparison tests (e.g., PISA; OECD, 2019) and are similar in many respects: both are small northern European Union nations, with well-educated workforces and similar language roots (Kikas & Lerkkanen, 2011). Children in both countries start school at the age of 7. Schools in both countries adhere to a prescribed national curriculum. However, there is a difference in the timing of formal literacy instruction, which starts in kindergarten in Estonia but in the first grade in Finland (Soodla et al., 2015). There are also differences in literacy instruction after the third grade. In Finland, the same teacher is responsible for all academic subjects across Grades 1–6; this teacher often is responsible for the same group of students for all 6 years. Literacy learning at Grades 1–3 emphasises reading accuracy, fluency, basic reading comprehension and spelling skills; at Grades 4–6, the emphasis shifts to independent reading, complex text comprehension and text production. From Grades 1 and 2 to Grades 3–6, literacy lessons decrease from 7 to 4–5 h/week. In Estonia, Grades 4–6 constitute the second stage of basic school. As in Finland, the same teacher is responsible for all main subjects in Grades 1–3. From Grade 4 onward, Estonian-language teachers are responsible for language-literacy lessons. Language-literature lessons are integrated in Grades 1–3 (19 lessons per week); in Grades 4–6, Estonian language (11 lessons per week) and literature (4 lessons per week) are taught separately. Thus, less attention is given to fluency than to text comprehension. Students in Estonia take a national-level language test at the close of the sixth grade, for which teachers prepare students intensively. Despite the change of curriculum foci and teaching practices during the late primary school years, early teaching practices are thought to affect later reading development by laying the foundation of reading skills (Snow & Matthews, 2016; Soodla et al., 2015), reading habits (Torppa et al., 2020) and motivation (Kikas & Tang, 2019; Lerkkanen et al., 2016). The present study The present study will address three research questions. First, are the same five teaching styles present in Finland and Estonia? Our previous study (Tang, Kikas, et al., 2017) reported five teaching styles among Finnish and Estonian third-grade teachers, yet they were identified using a combined sample. We expected to see same teaching styles within each country given similarities in their educational systems. Although there are differences in literacy instruction between two countries, in general, there are more similarities than differences (Kikas & Lerkkanen, 2011). In addition, teaching styles and their underlying teaching practices are broader than the literacy instruction; they represent more general pedagogical approaches across subject domains. TANG, KIKAS, PAKARINEN, LAURSEN & LERKKANEN © 2022 UKLA. Second, are third-grade teaching styles associated with changes in student reading skills from the third to sixth grade? On the basis of previous findings (Block, Parris, Reed, Whiteley & Cleveland, 2009; Tang et al., 2019; Tang, Kikas, et al., 2017), we hypothesised that the child-centred style and the mixed teaching style would be more strongly associated with reading fluency and reading comprehension than the child-dominated style. We also hypothesised that the teacher-directed style would be more strongly associated with reading fluency (Stipek, Feiler, Daniels & Milburn, 1995) than the child-dominated style. Third, are similar longitudinal associations between teaching styles and reading development found in both two countries? On the basis of previous research (Kikas & Lerkkanen, 2011; Tang, Kikas, et al., 2017), we expected the same pattern of associations in Finland and Estonia. Method Participants The sample included 70 third-grade teachers (33 in Finland and 37 in Estonia) and their students (N= 1,057: N Finland = 456 and N Estonia = 601). Most of the teachers were female (Finland = 72% and Estonia = 100%) and had at least an MA degree (Finland = 97% and Estonia = 100%). The modal length of teaching experience was 11 to 15 years in Finland and more than 15 years in Estonia. Most children were between the age of 9 and 10 years at the outset (Finland = 9.15 years, SD = 0.3; Estonia = 9.53 years, SD = 0.52). Boys comprised 48.7% of the Finnish students and 52.6% of the Estonian students. Procedure. Data were collected in classes during regular school hours by trained research assistants. The third-grade data were collected in the spring of 2010 (Finland and Estonia) and 2011 (three classes in Estonia) academic years; the sixth-grade data were collected in the spring term 3 years later. Teaching practices were observed during the third-grade data collection period. Detailed information on the teaching observation, coding process and inter-rater reliabilities is described in an earlier publication (Tang, Pakarinen, et al., 2017). Measures Classroom observations. Early Childhood Classroom Observation Measure was used to measure the degree (i.e., proportion of time) to which teaching practices in a classroom were child-centred, teacher-directed and child-dominated. Each practice was rated on three subscales: management (four items; e.g., children are allowed to take responsibility to the degree that they are able), climate (four items; e.g., teacher encourages children to engage in conversation and elaborate on their thoughts) and instruction (six items; e.g., tasks and lessons are designed to teach identifiable concepts and develop understanding). The rating scale was based on the percentage of time that each practice was demonstrated during the observation period: 1 = practice is rarely seen (0–20% of the time) to 5 = practice predominates (80–100% of the time). The measure has been validated in Finnish and Estonian educational contexts (Lerkkanen et al., 2012; Tang, Pakarinen, et al., 2017). The present study utilised a summary score of 14 items to assess each teaching practice. TEACHING STYLE AND FOLLOW-UP READING © 2022 UKLA. Reading fluency. A word-level reading fluency test was administered (Lindeman, 1998). In this speed test, a maximum of 80 items could be attempted within a 2-minute time limit. For each item, children selected one of four phonologically similar responses that matched the picture presented. Scores represent the number of correct answers (maximum = 80). Test-rest reliability was adequate (>.60). Reading comprehension. A reading comprehension test (Lindeman, 1998) assessed skills in gleaning factual knowledge, concepts and inferences from text. After silently reading the text, children were presented with 12 multiple-choice questions about the content. Scores represent the number of correct answers (maximum = 12). This normed test included different text and multiple-choice questions for each grade. Internal reliability was acceptable across countries and grades (Cronbach’s alpha: Finland = .66 and Estonia = .70–.77). All reading skills test scores were standardised within country. Missing data Of 1,057 students participating in the third grade, 722 (68.3%) also participated in the sixth grade. Little’s missing completely at random test indicated that data were not missing completely at random (χ 2 = 55.274, df = 19, p<.001). There were no differences as a function of age, country, gender or maternal educational, but students who participated at both grades had better reading fluency (t= 3.09, Cohen’sd= .20, p<.01) and reading comprehension (t= 2.62, Cohen’sd= .17, p<.01) than students who participated in the third grade only. To maximise statistical power and avoid bias arising from listwise deletion, multiple imputation was applied used to the missing values (Enders, 2013). Five data sets were generated, each with 50 iterations (Rpackage mice; van Buuren & Groothuis-Oudshoorn, 2011). Analysis strategy To determine whether teaching-style profiles replicated across countries, identical profile analyses were conducted on the data from Finland and Estonia. Following guidelines by Morin, Meyer, Creusier and Biétry (2016), we conducted latent profile analyses to estimate the number of profiles for each country and then contrasted profiles across countries. The first step was to establish configural similarity by determining whether the same number of profiles can be identified across two countries. The second step involved comparisons of structural similarity, constraining the within-profile means to be equal across countries. Dispersion similarity was then estimated by additionally constraining the within-profile variation to be equal across two countries. Finally, distributional similarity was examined by constraining the size (class proportion) of profiles to be equal across countries. Models were compared with Bayesian information criterion, consistent Akaike’s information criterion and sample size-adjusted Bayesian information criterion, with lower values indicating better model fit. Other model comparison indicators, such as bootstrapped likelihood ratio test and entropy, were also considered (see the supporting information for the evaluation and selection of models). We chose the default model specification on variances and covariances in latent profile analyses (i.e., equal variances and covariances fixed to 0). The analyses were conducted with MPLUS V8.3 (Muthén & Muthén, 2019). TANG, KIKAS, PAKARINEN, LAURSEN & LERKKANEN © 2022 UKLA. To test the hypothesis that third-grade teaching styles predicted changes in reading skills from the third to sixth grade, we conducted analyses of covariance (ANCOVAs) with teaching styles and country as independent variables, and sixth-grade reading fluency and comprehension as dependent variables. Age, gender, maternal education, and third-grade reading fluency and comprehension were included as covariates. ANCOVAs were conducted through Rpackage mitml (Grund, Robitzsch & Luedtke, 2019) and multcomp (Hothorn, Bretz & Westfall, 2008). Because ANCOVAs from multiple imputed data sets can only be obtained by using the parameters and covariance matrices of a regression analysis (van Ginkel & Kroonenberg, 2014), the interpretation of results differs from that of a standard ANCOVA because pooled Ftests must be used. All the supporting information can be accessed from https://osf.io/zmjv5/. Table 1. Associations between third-grade teaching styles and changes in reading skills from the third to sixth grade: pooled analysis of covariance estimates. Grade 6 reading fluency Grade 6 reading comprehension Estimate Std. error Estimate Std. error Intercept .81 0.57 .41 0.49 Grade 3 teaching style (ref. mixed style) Teacher-directed .06 0.19 .23 0.24 Extreme child-centred .38* 0.18 .01 0.23 Child-centred .09 0.09 .04 0.12 Child-dominated .02 0.29 .26 0.28 Country a .26** 0.10 .11 0.11 Age .09 0.08 .06 0.07 Gender b .18* 0.07 .24*** 0.06 Maternal education c Secondary .12 0.14 .01 0.11 More than secondary .07 0.12 .18 0.12 Grade 3 reading fluency .57*** 0.03 .17*** 0.04 Grade 3 reading comprehension .07 0.04 .27*** 0.04 Teaching Style × Country d Teacher-directed d .16 0.23 .10 0.25 Extreme child-centred d .06 0.21 .07 0.26 Child-centred d .28 0.16 .14 0.16 Child-dominated d .26 0.28 .23 0.35 a Finland as the reference group; thus, positive estimate means Estonia is greater than Finland. b Girl as the reference group. c Less than secondary as the reference group. d Estonia and mixed teaching style as the reference group. * p<.05. ** p<.01. *** p<.001. TEACHING STYLE AND FOLLOW-UP READING © 2022 UKLA. Results Identification of teaching styles in Finland and Estonia Latent profile analysis confirmed that five profiles represented the best solution in each country (Table S1). A mixed teaching style,extreme child-centred teaching style, child-centred teaching style,teacher-directed teaching style and child-dominated teaching style were identified in Finland and Estonia. Multiple-group latent profile analysis revealed dispersion similarity (i.e., invariant means and variance), indicating that five teaching styles were equivalent across countries (Table S1). Teaching styles and changes in reading performance from the third to sixth grade The role of third-grade teaching styles in reading development from the third to sixth grade was tested using pooled ANCOVAs (see Table 1 for findings using the mixed teaching style as the reference group; see Table 2 for findings using other reference groups). Comparisons with the mixed teaching style (Table 1) indicate that the extreme child-centred teaching style was a less effective practice for the development of reading fluency (estimate = .38, p<.05). The child-centred teaching style was more effective than the extreme child-centred teaching style in terms of reading fluency (estimate = .47, p<.05; Table 2). The extreme child-centred teaching style was marginally less effective than the teacher-directed style in reading fluency (estimate = .44, p= .089). There was a main effect of country on reading fluency (Table 1). Estonian students scored higher on reading fluency than their Finnish peers (estimate = .26, p<.01). Boys scored lower on reading fluency than girls (estimate = .18, p<.05). There were no main effects for child age or maternal education. Finally, there were no two-way interactions between country and teaching styles on reading fluency. There were neither main effects nor interactions for any study variable on reading comprehension (Tables 1 and 2). Table 2. Pooled analyses of covariance of other teaching styles as reference group. Reading fluency Reading comprehension Estimate Std. error Estimate Std. error Extreme child-centred vs teacher-directed .44 0.26 .24 0.33 Child-centred vs teacher-directed .03 0.20 .28 0.22 Child-dominated vs teacher-directed .08 0.34 .49 0.34 Child-centred vs extreme child-centred .47* 0.19 .04 0.24 Child-dominated vs extreme child-centred .35 0.34 .25 0.31 Child-dominated vs child-centred .12 0.27 .21 0.27 * p<.05. ** p<.01. *** p<.001. TANG, KIKAS, PAKARINEN, LAURSEN & LERKKANEN © 2022 UKLA.