Learning and Individual Differences 23 (2013) 72–79
Contents lists available at SciVerse ScienceDirect
Learning and Individual Differences journal homepage: www.elsevier.com/locate/lindif
Teachers adapt their instruction in reading according to individual children's literacy skills☆ Jari-Erik Nurmi a,⁎, Noona Kiuru a, Marja-Kristiina Lerkkanen b, Pekka Niemi d, Anna-Maija Poikkeus b, Timo Ahonen a, Esko Leskinen c, Anna-Liisa Lyyra a a
Department of Psychology, University of Jyväskylä, Finland Department of Teacher Education, University of Jyväskylä, Finland Department of Mathematics and Statistics, University of Jyväskylä, Finland d Department of Psychology, University of Turku, Finland b c
a r t i c l e
i n f o
Article history: Received 26 August 2011 Received in revised form 21 May 2012 Accepted 16 July 2012 Keywords: Adaptive teaching Evocative impact Instruction Literacy Students Teachers
a b s t r a c t This study examined the extent to which first grade teachers adapt their reading instruction to the literacy skills of particular children in their classroom, and investigated whether teacher and classroom characteristics influence such adaptation. Three hundred seven Finnish children were tested with regard to their literacy skills at the end of their kindergarten year. At the beginning of the first grade, the teachers of these children filled in a questionnaire on the reading support they had given each child. The results showed, first, that the poorer the literacy skills a child showed at the end of kindergarten, the more personal reading instruction the teacher gave the child in grade 1. Second, teachers who were less experienced or had fewer teaching assistants in the classroom adapted their instruction on the basis of children's literacy skills to a greater extent than other teachers did. © 2012 Elsevier Inc. All rights reserved.
1. Introduction Although previous research has shown that adapting reading instruction to students' literacy skills has a positive impact on their skill development (Connor, Morrison, Fishman, Schatschneider, & Underwood, 2007; Connor, Morrison, & Katch, 2004), only few attempts have been made to directly investigate the extent to which a student's academic skills influence his or her teacher's instruction for that particular student (Pianta, 2006). Many teacher and classroom characteristics—such as professional experience (e.g., Connor, Son, Hindman, & Morrison, 2005; Kontos & Wilcox-Herzog, 1997; Pianta et al., 2005), class size (e.g., Blatchford, 2003; NICHD, 2004), and the number of teaching assistants in a classroom—may also influence the extent to which teachers adapt their instruction according to particular students' skills. To investigate these assumptions, the present study addressed three specific questions: First, to what extent do first grade teachers adapt their reading-related instruction to the literacy skills of a particular child in
☆ The work presented in this article was funded by the Academy of Finland (Grants 213486, 126043, 7133146 and 7133146). ⁎ Corresponding author at: Department of Psychology, P.O. Box 35, 40014 University of Jyväskylä, Finland. Tel.: +358 504285308; fax: +358 142602200. E-mail address: jari-erik.nurmi@jyu.fi (J.-E. Nurmi). 1041-6080/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.lindif.2012.07.012
their class? Second, to what extent do teachers differ in the extent to which they adapt their reading-related instruction? Third, to what extent do structural characteristics in classroom teaching—that is, the teacher's teaching experience, the number of teaching assistants, and the number of students in the class (i.e., class size)—moderate the association between a child's literacy skills and the instruction the teacher provides to him or her in the class? 1.1. The role of students' academic skills in shaping teachers' instruction The notion that a particular student's academic performance activates a certain kind of instruction has been discussed by many researchers (Babad, 1998; Corno, 2008; Pianta, 2006; Skinner & Belmont, 1993) and can be described as individual students' “evocative impact” on teachers' instructional practice (Rutter, 1997; Scarr & McCartney, 1983). Originally, this idea of a child's impact on his or her interpersonal environment was introduced in family studies (Bell, 1968; Hartup & Laursen, 1991). There are at least two psychological mechanisms that may be responsible for how students' characteristics impact their teachers' behavior and instruction. First, teachers' cognitive mechanisms are involved. Besides having general pedagogical knowledge concerning classroom management and subject matters, teachers also construct knowledge with regard to particular students (Borko & Putnam, 1996; Calderhead, 1996). By constructing student-related knowledge,
J.-E. Nurmi et al. / Learning and Individual Differences 23 (2013) 72–79
teachers are able to plan their teaching tasks and can design appropriate instructional content (e.g., instruction that is code-focused vs. meaning-focused and teacher–child oriented vs. child-managed; Calderhead, 1996; Connor et al., 2009). Second, students and their characteristics also activate affective reactions among their teachers. Some studies suggest that the major source of teachers' positive emotions, such as satisfaction, are children's successful learning outcomes and progress (Emmer, Oakland, & Good, 1974; Hargreaves, 1998). Teachers' emotions may then also have important consequences for the kind and amount of instruction they direct at a particular student in their class (Sutton & Wheatley, 2003). Despite researchers' interest in the role played by an individual students' academic performance and the ways in which teachers deal with that student, our understanding of the phenomenon is limited (Doyle, 1981; Pianta, 2006). The major reason for this is that previous research on the topic is scarce and the findings are contradictory (Nurmi, in press). Early studies in the field, focusing on teachers' expectations of students, showed that high achieving students received more preferential treatment from their teachers, such as higher academic demands and privileges, whereas low achieving students received more criticism, concern, and management (Good & Brophy, 1973; Weinstein & Middlestadt, 1979). Some of the more recent studies have found that, although high achieving students receive more emotional support and positive affect from their teachers, low achieving students receive more instructional support and active instruction (Babad, 1990, 1996; Nurmi, Viljaranta, Tolvanen, & Aunola, 2012). Consequently, the first aim of the present study was to examine to what extent teachers adapt the reading instruction they give to a particular child at the beginning of the first grade according to the child's current literacy skills. In the present study, the reading-related instruction received by individual children was assessed using teachers' reports concerning the extent to which they had provided support for and gave attention to particular children with respect to learning to read. 1.2. Teacher differences in the adaptation of instruction according to students' academic skills While it has been shown that the instruction given by teachers, such as in reading, varies in many ways (Connor et al., 2004, 2009; Juel & Minden-Cupp, 2000), teachers may also differ in the extent to which they adapt their instruction according to individual students' academic skills. The differences between teachers in such adaptation may be influenced by a variety of teacher and classroom characteristics. For example, teachers with less teaching experience have been shown to be more influenced by contextual factors (Tschannen-Moran & Woolfolk Hoy, 2007), such as students' level of academic performance, than those with more teaching experience. Moreover, as teachers with less experience tend to show a higher emotional quality toward students than teachers with more experience tend to do (NICHD, 2004), such emotional quality may be important in activating attention toward a particular student who exhibits problems in his or her performance (Sutton & Wheatley, 2003). The differences between teachers in their instructional adaptation to individual students' skills may also be influenced by the number of teaching assistants present in the classroom at the same time. On the one hand, the teaching support of additional adults in the classroom may give the teacher greater freedom to adapt his or her instruction. On the other hand, when the teacher does not have an assistant in the classroom, he/she may be obliged to focus more on the students who are doing poorly. Also, class size (i.e., the number of students in the classroom) may influence the differences between teachers regarding the amount of adaptation they apply to a particular student's level of academic performance. For example, teachers of smaller classes may have more resources and time to adjust the instruction they give a particular student than do teachers of larger classes.
73
Previous studies that have investigated the influence of class size, teachers' teaching experience, and the number of teaching assistants on class-level instruction have shown mixed findings. For instance, most studies on the role of teachers' teaching experience in the quality of instruction have not found any associations (Henninger, Flowers, & Councill, 2006; Myrberg, 2007; Stipek & Byler, 2004). However, in a study by Connor et al. (2005), teachers with more years of experience were less warm and responsive to their students than were teachers with fewer years of experience. The extent to which teaching assistants contribute to classroom instruction has seldom been investigated. In one study, a minor association between the number of teaching assistants and a positive emotional composite in classroom interactions was found (NICHD, 2002). Moreover, some studies on class size have shown that smaller classes result in a higher quality of instructional and emotional support (Blatchford, 2003; NICHD, 2004). As far as we know, no research results have been published on the extent to which classes' structural variables contribute to how far teachers adapt their instruction in reaction to their students' academic skills. The present study was carried out in Finland. Finnish schools differ in at least three important ways from many other educational settings. First, the children are from one to two years older when they start primary school than children in many other countries. In Finland, a child's compulsory education begins in the year in which the child turns seven years of age. Six-year-olds are entitled to kindergarten education for one year before embarking on their nine-year compulsory school career (National Board of Education, 2008). Second, kindergarten education in Finland focuses on learning through play, participation and creativity, and systematic instruction in reading starts only when children move to primary school. Although the curriculum covers seven subject areas, lessons are not divided into individual subjects as such. Third, compared to many orthographically opaque languages (e.g., English and French; see, Seymour, Aro, & Erskine, 2003), learning to read Finnish is relatively easy since it has a highly regular orthography and simple syllabic structure. Instruction in reading leans heavily on phonics, enabling beginning readers to decode new written words by blending sound-spelling patterns. One should also keep in mind that the classrooms examined as part of the present study represent one particular cultural setting with specific historical and political precedents for particular practices.
1.3. Aims of the study The present study investigated the following research questions: (1) To what extent do 1st grade teachers adapt their reading instruction according to a particular child's literacy skills (measured in the spring of the kindergarten year)? Our hypothesis (1) was that teachers focus more on instructing students who show poor skills in literacy than on students who demonstrate good skills (Corno, 2008; Pressley, Hogan, Wharton-McDonald, Mistretta, & Ettenberger, 1996). (2) To what extent do 1st grade teachers differ in the extent to which they adapt their reading instruction according to the literacy skills of a particular child? (2) We hypothesized that teachers do differ in the extent to which children' literacy skills influence the support in reading they provide to particular children (Connor et al., 2009; Pakarinen, Lerkkanen, Poikkeus, Siekkinen, & Nurmi, 2011). (3) To what extent do classes' structural characteristics—i.e., teachers' teaching experience, the number of teaching assistants, and class size—moderate the impact of children's literacy skills on the instruction teachers give a particular child. We set several hypotheses concerning this point:
74
J.-E. Nurmi et al. / Learning and Individual Differences 23 (2013) 72–79
(3a) We hypothesized, first, that teachers with less teaching experience adjust their instruction to a greater extent than do teachers with more teaching experience, because they have been found to be warmer and more responsive to their students than teachers with more teaching experience (Connor et al., 2005), and they have been trained more recently and may consequently have learned more child-centered strategies (Hytönen, 2008). Two alternative hypotheses were set regarding the number of teaching assistants. (3b) As teachers in classrooms with additional adult help can be assumed to have more freedom in choosing their classroom activities, we assumed that the more teaching assistants are in the classroom, the more the teacher adapts his/her instruction according to particular children's literacy skills. (3c) Alternatively, when teachers do not have assistants in the classroom, they may be obliged to adapt their instruction for particular students who are performing especially poorly as regards literacy. (3d) As previous research has shown that smaller classes tend to result in a higher quality of instructional support (Blatchford, 2003; NICHD, 2004) than is the case for larger classes, we assumed that teachers of smaller classes adapt the instruction they give a particular student according to that student's academic skills to a greater extent than do teachers of larger classes. 2. Method 2.1. Participants 2.1.1. Children The participants comprising the sample of our study were 307 (151 girls, 156 boys) children who were studied from kindergarten to primary school as part of the more extensive First Steps study (Poikkeus et al., 2006). In the fall of their kindergarten year, the majority of the children were six years old (M = 74.0 months, SD = 3.4 months). The children came from kindergartens and primary schools that were situated in three medium-sized towns and one municipality in Finland. Originally, 361 target children (179 girls, 182 boys) were randomly selected from a larger sample of 1 880 children (upon kindergarten entry) who were participating in the First Steps study. The aim of the sampling was to randomly select three students from each classroom. Due to some either particularly small or large class sizes, the actual number of children selected from the different classrooms ranged between one and six. The reason for creating the subsample was to decrease teachers' workload. Due to the fact that of the 139 teachers originally contacted, 22 did not provide ratings of the students in their classroom, the final sample of students in the present study was 307. Parental consent was requested and received for all 307 children. The children's literacy skills were examined in the spring of their kindergarten year (T1). 2.1.2. Teachers Teachers were asked for their written consent to be entered in the study. Of the 139 teachers contacted, one hundred seventeen first grade teachers from 117 classrooms in 110 schools agreed to participate in the study. Forty-five percent of the teachers had more than 15 years of teaching experience, 36% had from 6 to 15 years experience, 15% had 1 to 5 years experience, and 4% had less than one year of experience (Mode = more than 15 years). In the fall of the first grade year (T2; October), 117 teachers (107 females, 10 males) rated the extent to which they supported each child in reading. At this time point, the teachers had taught the children from 6 to 8 weeks. In the spring of the first grade year (T2) 110 of these 117 teachers filled in a questionnaire on their teaching
experience, class size, and the number of teaching assistants in the classroom. 2.2. Measures 2.2.1. Children's literacy skills in the spring of kindergarten (T1) All literacy skills were individually assessed in the spring of the kindergarten year (T1). 2.2.1.1. Phoneme identification. Initial phoneme identification was tested using the ARMI test (Lerkkanen, Poikkeus, & Ketonen, 2006). The children were shown four pictures of objects with the matching names. The children were then asked to select the correct picture on the basis of the spoken initial phoneme of the name of the target object (e.g. At the beginning of which word do you hear ____?). The sum score was based on the number of correctly answered items (maximum score 10). The Cronbach's alpha for Phoneme Identification was .73. 2.2.1.2. Letter knowledge. Children were asked to name 29 uppercase letters shown by the examiner. The letters were in random order, arranged into three rows and shown one row at a time (subtest of the ARMI; Lerkkanen et al., 2006). Uppercase letters were used since the children in Finnish kindergartens are mainly exposed to letters in this form. The use of a phoneme or a letter name was coded as a correct response. The child received one point for each correct response (maximum score = 29). The Cronbach's alpha for Letter Knowledge was .94. 2.2.1.3. Word reading. Reading of words was assessed using an individually administered word list (subtest of the ARMI; Lerkkanen et al., 2006). The list contained 10 words. The words were two-syllable (7 words), three-syllable (2 words) and five-syllable (1 word) words. The sum score of the correct items was calculated with regard to each child's performance. The Cronbach's alpha for Word Reading was .86. 2.2.1.4. Text reading fluency. Text reading fluency was assessed using an oral reading test. The text Jännittävät Matkat (meaning Exciting Adventures), comprised of 124 words, was used to assess text reading fluency. The fluency score was calculated by counting the number of words correctly read within one minute (Lyytinen, Eklund, & Lyytinen, 2005). The text reading test was only administered for those children who scored 9 or 10 points on the Word Reading test. Other children (nonreaders) were assigned the score zero on this test. The test-retest correlation over the period of half a year was .87 for this test. The composite score for Literacy Skills was calculated as follows. First, the scales of the children's tests scores on the Phoneme Identification, Letter Knowledge, Word Reading, and Text Reading Fluency tests were standardized. Then, a composite mean sum score was calculated based on the four test scores. The Cronbach's alpha for the Literacy Skills composite score was .81. 2.2.2. Measures of first grade teachers 2.2.2.1. Perceived teacher support in reading. In October (T2) of the first grade, the teachers were asked to rate, on a five-point scale, the extent to which they provided special support for and gave extra attention to particular children with respect to learning to read during classroom teaching hours in comparison to most of the other children in the class (i.e., the relative amount of reading support given). Additional analyses showed that perceived teacher support in reading in the fall of the 1st grade year, and the same variable measured in spring, showed a correlation of .69 (p b .001). Moreover, perceived teacher support in reading correlated .86 (p b .001) with teacher
J.-E. Nurmi et al. / Learning and Individual Differences 23 (2013) 72–79
support in writing, in this sample. The data of the subsample of 97 children with a risk for reading disability showed that the instruction reported by the classroom teachers correlated .48 (p b .001) with the amount of remedial reading instruction in grade 1 reported by special education teachers. 2.2.2.2. Background questionnaire. Teachers were first asked to report the length of their experience of working as a primary school teacher, using the following six alternatives: 0=not at all, 1=less than a year, 2= 1–5 years, 3=6–10 years, 4=11–15 years, 5=more than 15 years. A variable measuring less vs. more teaching experience was dummycoded in the following way: 0=less than 15 years of teaching experience (55%), 1=more than 15 years of teaching experience (45%). This recoding was undertaken to distinguish those teachers who had gone through the old type of teacher education from those who went through the more recent teacher education program (Kansanen, 2008). Teachers were also asked how many teaching assistants they had in their classroom. The answer to this question was coded in the following way: 0 = no assistant (30%), 1 = one assistant (60%), 2 = 2 assistants (7%), and 3 = 3 or more assistants (2%). Finally, teachers were asked to report the size of their class (M = 17.5; SD = 5.90). 2.3. Analysis strategy The first research question was examined by estimating the path leading from students' literacy skills in kindergarten (T1) to the level of reading support that teachers gave to particular children in the first grade (T2), using the Complex approach (Muthén & Muthén, 1998–2008). Instead of multilevel modeling, this method estimates the model at the level of the whole sample but corrects distortions of standard errors in the estimation that may have been caused by the clustering of observations (Fig. 1a). The following regression equation corresponds to Fig. 1a: 2 εeN 0; σ ε ;
y ¼ β0 þβ1 x þ ε;
where β0 is a fixed intercept and β1 is a fixed regression coefficient. Next, multilevel modeling with random slopes (e.g., Muthén & Muthén, 1998–2008; Raudenbush & Bryk, 2002; Snijders & Bosker, 1999) was used to examine whether the strength of the association between children's literacy skills and the reading support for particular children varied among the teachers (see Fig. 1b; classroom level above the dashed line and individual child level below the dashed line), and whether some classroom-level variables predict this variation (see Fig. 1c). The following equations correspond to Fig. 1b: Level-1 (within-level, within classrooms) y ¼ i þ sðxcw Þ þ ε;
2 εeN 0; σ ε ;
where xcw ¼ x−x is a group-mean centered covariate at the within-level (literacy skills in kindergarten),i is a random intercept term, and s is a random slope term (see Fig. 1b). Level-2 (between-level, between classrooms) i ¼ α0 þ ζ 0 ; s ¼ α1 þ ζ 1 ;
ζ 0 eN 0; σ 2i ; 2 ζ 1 eN 0; σ s :
In the equations, i refers to the random intercept and s refers to the random regression coefficient (Fig. 1b). At the individual level of
75
this model (Fig. 1b), perceived teacher support in reading in grade 1 (T2) is predicted by students' literacy skills in kindergarten (T1). The individual-level predictor (i.e., literacy skills) is group-mean centered. We were particularly interested in investigating whether statistically significant variation exists between classrooms with regard to the strength of the slope s, i.e., the regression coefficient pointing from literacy skills to the teacher support in reading. Next, we examined whether some classroom-level variables predict this variation (see Fig. 1c). The equations presented in Fig. 1c were the same as those in Fig. 1b, except that the random slope s was predicted by the classroom-level variables. The equations of Level-2 were as follows: Level-2 (between-level, between classrooms) i ¼ α0 þ ζ 0 ; ζ 0 eN 0; σ 2i ; s ¼ α 1 þ γ 1 xcb1 þγ 2 xcb2 þγ 3 xcb3 þζ 1 ;
2 ζ 1 eN 0; σ ς1
In the between-level equations, γ-parameters refer to the regression coefficients leading from the three grand-mean centered between-level covariates to the random slope (Fig. 1c). All of the analyses were carried out with the Mplus statistical package (version 5.1; Muthén & Muthén, 1998–2008), using the missing data method, that is, the standard missing-at-random (MAR) approach to missingness (Muthén & Muthén, 1998–2008). This missing data method uses all of the relevant data available in order to estimate a model without imputing data. The parameters of the models were estimated using the MLR estimator (Muthén & Muthén, 1998–2008). 3. Results We used the Complex method (Muthén & Muthén, 1998–2008) to examine whether literacy skills in kindergarten (T1) would predict individual instruction in reading given to a particular child in the fall of the first grade year (T2). The results showed that the poorer the students' literacy skills were at the end of kindergarten, the more individual instruction they received in reading at the beginning of their ^ ðs:eÞ ¼ 0:64ð0:04Þ; pb:001; R2 ¼ :41). first school year (Standardizedβ Next, the intraclass correlation (Heck, 2001; Muthén, 1991) for literacy skills variable was calculated. The results showed that 9% of the total variance in literacy skills in kindergarten spring was due to differences between classes in the first grade, confirming that classes differed with respect to children's skill levels already before the first grade. The results of the multilevel models with random slopes showed further that the strength of the regression coefficient pointing from kindergarten literacy skills to perceived teacher support in reading in grade 1 varied between classes ( σ^ 2s ðs:eÞ ¼ 0:33ð0:17Þ; p b .01; one-tailed test, H1: between-variance is larger than zero; CI (95%) = [0.06, 0.61]): teachers differed in the extent to which they adapted their reading-related instruction according to individual students' literacy skills. There was also statistically significant classroom-level variation in the overall amount of reading support given by teachers to the students in the class. ( σ^ 2i ðs:eÞ ¼ 0:41ð0:09Þ; pb:001; one-tailed test, H1: betweenvariance is larger than zero; CI (95%)=[0.26, 0.57]). However, as our measure for reading instruction measures the relative amount of instruction per child, an interpretation of classroom-level variation in the overall instruction is not informative. To examine whether classes' structural characteristics would moderate the impact of students' literacy skills on perceived teacher support in reading, the random slope at the between-level was predicted simultaneously by teaching experience, the number of teaching assistants, and class size (Table 1). The results showed (Fig. 2) that the less teaching experience the teacher had, the larger was the negative association between students' literacy skills and the individual instruction the teacher provided. Also, the fewer assistants the teacher had in the classroom, the stronger was
76
J.-E. Nurmi et al. / Learning and Individual Differences 23 (2013) 72–79
a
Teacher support in reading (T2, grade 1)
Literacy skills (T1, kindergarten)
y
b
1
1
s
(
2 s
)
1
0
i 2
(
i
)
Between Within Literacy skills (T1, kindergarten)
Teacher support in reading (T2, grade 1)
s
i
xc w
y
c xcb1
Teachers’ experience
1 1
xcb2
γ1
Teaching assistants
s
γ2
1
γ3
xcb3
Class size
i
0
0
Between-level Between Within-level Within Literacy skills (T1, kindergarten)
xcw
1
Teacher support in reading (T2, grade 1)
s i
y
Fig. 1. Theoretical models. a. Theoretical regression model for perceived teacher support in reading using the Complex approach. b. Theoretical random slope multilevel model without for perceived teacher support in reading. c. Theoretical random slope multilevel model with predictors for perceived teacher support in reading. Note. The within-level covariate is group-mean centered (xcw) and between-level covariates are grand-mean centered (xcb).
the negative association between students' literacy skills and the individual instruction the teacher provided as shown in Fig. 3 (3a and 3b). 4. Discussion Only few previous attempts have been made to investigate the influence of a student's academic skills on the instruction given by his or her teacher, or how different teacher and classroom characteristics explain differences between teachers concerning this type of influence. The results of the present study showed, first, that the poorer the performance in literacy skills a student showed at the end of
kindergarten, the more reading instruction the teacher gave to that particular child in first grade. Second, the less teaching experience and the fewer teaching assistants in the classroom, the more students' performance in reading influenced the instruction the teacher gave them. 4.1. The role of students' literacy skills in teachers' instruction Our first research question asked to what extent first grade teachers adapt their reading instruction according to the literacy skills of a particular child. As was expected (Hypothesis 1), the results
J.-E. Nurmi et al. / Learning and Individual Differences 23 (2013) 72–79 Table 1 Unstandardized estimates (standard errors in parentheses) of random slope components: multilevel model of perceived teacher support in reading in first grade (N between = 117, Nwithin = 306). Estimate (S.E.) Within-level (individual level) Residual variance (σ2ε) Between-level (classroom level) Intercept Regression coefficient from literacy skills (T1) to teacher support (T2, s) Regression coefficientsa s on teachers' length of teaching experience (γ1) s on number of assistant teaching personnel (γ2) s on class size (γ3) Residual variance Residual variance (s)
77
of primary school. By using a multidimensional model of reading instruction (code-focused vs. meaning-focused instruction, teacher– child vs. child-managed instruction, and change over time), Connor et al. (2009) showed that it was the precise amount of model-based individualized instruction that predicted students' good reading outcomes (see also Connor et al., 2004).
0.61(0.09)***
4.2. Teacher and classroom characteristics as moderators of the association between particular students' literacy skills and the teacher's instruction −1.18(0.11)*** 0.36(0.19)* 0.61(0.13)*** −0.003(0.02) 0.20(0.14)
Note. *** p b .001; ** p b .01; * pb .05. a Between‐level predictors were allowed to correlate.
showed that a student's literacy skills at the end of kindergarten were a powerful predictor of the individual reading support teachers gave a particular student (41% of variance): the poorer the performance in literacy skills a student showed, the more reading support the first grade teacher gave to that particular student. Our finding is in line with those of some previous studies showing that low-achieving students receive more support and active instruction than high-achieving students do (Babad, 1990, 1996; Babad, Bernieri, & Rosenthal, 1989; Nurmi et al., 2012). Overall, our findings lend support to the classroom instruction theories according to which teachers typically adjust their instruction to students' characteristics, such as in reaction to students' level of academic skills (Corno, 2008; Doyle, 1981; Pianta, 2006). This is an important finding because it suggests that students' characteristics have an “evocative impact” on teacher–child interaction, similar to what has been shown regarding parent–child interaction (Bell, 1968; Hartup & Laursen, 1991; Pomerantz & Murry Eaton, 2001). As the results pertain to Finland, and thus to the context of one particular educational and cultural setting with specific historical and political precedents for particular practices, there is an evident need to replicate the findings in other cultural settings. Previous research has also shown that adapting (individualizing) reading instruction according to students' decoding skills and vocabulary improves students' literacy development during the first years
Our next research question enquired whether teachers differ in the extent to which they adapt their reading instruction according to a particular child's level of literacy skills, and whether teacher and classroom characteristics predict such differences. The results showed, first, that teachers differed in how far they adapted their class instruction for the sake of a particular student according to the child's literacy skills. The results showed further, as expected (Hypothesis 3a), that the less teaching experience teachers had, the more they adjusted the instruction they gave a particular student according to the student's level of literacy skills. A similar pattern was recently found by another study using a different methodology (Nurmi et al., 2012). There are at least three possible explanations for the findings of the present study. First, previous results have suggested that teachers with less experience show a higher emotional quality toward students than teachers with more experience tend to do (NICHD, 2004), which may be important in activating attention toward a particular student who exhibits problems in his or her performance (Sutton & Wheatley, 2003). Second, previous research has shown that the beliefs of teachers with less teaching experience are more influenced by contextual factors, that is, teaching resources and interpersonal support, than is true for teachers with more teaching experience (Tschannen-Moran & Woolfolk Hoy, 2007). In our study, in the case of teachers with less experience, a further contextual factor—students' level of academic performance—was found to be influential concerning teachers' instruction given to a particular student in their class. The final explanation relates to the changes that have taken place in teacher education (Babad, 1998). For example, before 1979, teacher training in Finland was a three-year program carried out in Teacher Training colleges and placed a heavy emphasis on didactics. In 1979, teacher training was reformed and a 4-year Master's degree program was introduced (Kansanen, 2008). Between 1979 and the early 1990s, the emphasis in teacher training gradually shifted away from didactics and toward educational psychology.
Teachers’ experience
.29* s (R = .49) 2
.59***
Teaching assistants
i Between-level Between Within-level Within Literacy skills (T1, kindergarten)
Teacher support in reading (T2, grade 1)
s i
Fig. 2. Final multilevel random slope model with predictors for teachers' reading instruction. The paths are presented as standardized estimates. * p b .05; ** p b .01; *** p b .001.
78
J.-E. Nurmi et al. / Learning and Individual Differences 23 (2013) 72–79
a Teacher support in reading in grade 1(z-score)
0,6 0,4
Less than 15 years of teaching experience
0,2 0
More than 15 years of teaching experience
-0,2 -0,4 -0,6 -0,8 Low literacy skills in kindergarten (-1 SD)
High literacy skills in kindergarten (+1 SD)
Literacy skills in kindergarten
b
1
Teacher support in reading in grade 1(z-score)
0,8 0,6 No teaching assistants
0,4 0,2
One teaching assistant
0 -0,2
Two or more teaching assistants
-0,4 -0,6 -0,8 -1
Low literacy skills in kindergarten (-1 SD)
High literacy skills in kindergarten (+1 SD)
Literacy skills in kindergarten Fig. 3. Plots of cross-level interactions. a. Teacher experience as a moderator of the extent of adapting instruction in reading according to individual students' literacy skills. b. Number of teaching assistants as a moderator of the extent of adapting instruction in reading according to individual students' literacy skills.
In accordance with our Hypothesis 3c (and contrary to Hypothesis 3b), our results showed that the fewer teaching assistants there were in the classroom, the more teachers adapted their instruction according to particular students' literacy skills. This finding suggests that when teachers do not have assistants in the classroom, they are obliged to focus particularly on students who are performing poorly whereas when they have assistants, they can more easily direct their instruction at the majority of students. Also, contrary to our expectations (Hypothesis 3d), our results showed that class size did not impact the extent to which teachers adapted the instruction they provided according to particular students' literacy skills (see also, NICHD, 2002; Rutter & Maughan, 2002). 4.3. Limitations The present study has some limitations that need to be taken into account when making generalizations on the basis of the findings.
First, our study focused on quantitative rather than qualitative aspects of teachers' instructional practices and contents of their reading instruction. Consequently, the qualitative differences in teachers' reading instruction that are activated by students' individual skill levels also merit attention in future research. One possibility for such analysis would be to collect ratings from teachers concerning their beliefs about adaptive teaching practices to see whether those beliefs connect with their actual performance. Second, the measure of teacher support in reading given to a particular child was based on the teachers' self-reports, and therefore our findings need to be replicated by using classroom observations. Third, our measure of perceived teacher support in reading given to a particular child was based on only one item. This was due to the fact that teachers were not willing to fill in too long a questionnaire for each child. Fourth, our study examined children when they were facing the transition from kindergarten to primary school. As it is possible that teachers
J.-E. Nurmi et al. / Learning and Individual Differences 23 (2013) 72–79
are particularly sensitive to students' level of academic skills during this transition, it would be important to also investigate the phenomenon of adaptive teaching with regard to the later school years. 4.4. Conclusion This study adds to previous research on teacher–child classroom interaction by showing that teachers adapt the reading instruction they give at the beginning of the first grade according to children's literacy skills. However, the less teaching experience teachers reported having and the fewer teaching assistants there were in the classroom, the more students' literacy skills contributed to the individual reading instruction they received from teachers. References Babad, E. (1990). Measuring and changing teachers' differential behavior as perceived by students and teachers. Journal of Educational Psychology, 82, 683–690. Babad, E. (1996). How high is “High inference”? Within-classroom differences in students' perceptions of classroom interaction. Journal of Classroom Interaction, 31, 1–9. Babad, E. (1998). Preferential affect: The crux of the teacher expectancy issue. In J. Brophy (Ed.), Advances in research on teaching: Expectations in the classroom, Vol. 7. (pp. 183–214)New York: JAI Press Inc. Babad, E., Bernieri, F., & Rosenthal, R. (1989). When less information is more informative: Diagnosing teacher expectations from brief samples of behavior. British Journal of Educational Psychology, 59, 281–295. Bell, R. Q. (1968). A reinterpretation of the direction of effects in studies of socialization. Psychological Review, 75, 81–95. Blatchford, P. (2003). A systematic observational study of teachers' and pupils' behaviour in large and small classes. Learning and Instruction, 13, 269–595. Borko, H., & Putnam, R. T. (1996). Learning to teach. In D. C. Berliner, & R. C. Calfee (Eds.), Handbook of Educational Psychology (pp. 673–708). New York: Simon & Schuster Macmillan. Calderhead, J. (1996). Teachers: Beliefs and knowledge. In D. C. Berliner, & R. C. Calfee (Eds.), Handbook of Educational Psychology (pp. 709–725). New York: Simon & Schuster Macmillan. Connor, C. M., Morrison, F. J., Fishman, B. J., Schatschneider, C., & Underwood, P. (2007). Algorithm-guided individualized reading instruction. Science, 315, 464–465. Connor, C. M., Morrison, F. J., & Katch, L. E. (2004). Beyond the reading wars: Exploring the effect of child-instruction interactions on growth in early reading. Scientific Studies of Reading, 8, 305–336. Connor, C. M., Piasta, S. B., Fishman, B., Glasney, S., Schatschneider, C., Crowe, E., et al. (2009). Individualizing student instruction precisely: Effects of child × instruction interactions on first graders' literacy development. Child Development, 80, 77–100. Connor, C. M., Son, S., Hindman, A., & Morrison, F. J. (2005). Teacher qualifications, classroom practices, family characteristics and preschool experience: Complex effects on first graders' vocabulary and early reading outcomes. Journal of School Psychology, 43, 343–375. Corno, L. (2008). On teaching adaptively. Educational Psychologist, 43, 161–173. Doyle, W. (1981). Research on classroom contexts. Journal of Teacher Education, 32, 3–6. Emmer, E., Oakland, T., & Good, T. (1974). Do pupils affect teachers' styles of instruction? Educational Leadership, 31, 700–704. Good, T. L., & Brophy, J. E. (1973). Behavioral expression of teacher attitudes. Journal of Educational Psychology, 63, 617–624. Hargreaves, A. (1998). The emotional practice of teaching. Teaching and Teacher Education, 14, 835–854. Hartup, W. W., & Laursen, B. (1991). Relationships as developmental contexts. In R. Cohen, & A. W. Siegel (Eds.), Context and Development (pp. 253–279). Hillsdale: Lawrence Erlbaum Associates. Heck, R. H. (2001). Multilevel modeling with SEM. In G. A. Marcoulides, & R. E. Schumacker (Eds.), New developments and techniques in structural equation modelling (pp. 89–127). Mahwah, NJ, US: Lawrence Erlbaum Associates. Henninger, J. C., Flowers, P. J., & Councill, K. H. (2006). Pedagogical techniques and student outcomes in applied instrumental lessons taught by experienced and pre-service American music teachers. International Journal of Music Education, 24, 71–84. Hytönen, J. (2008). Suomalaisen lapsikeskeisen pedagogiikan kehityslinjoja toisen maailmansodan jälkeen – eräitä kriittisiä havaintoja [The development of a child-centered pedagogy in Finland after the Second World War – Some critical observations]. Didacta Varia, 13, 19–25. Juel, C., & Minden-Cupp, C. (2000). Learning to read words: Linguistic units and instructional strategies. Reading Research Quarterly, 35, 458–492. Kansanen, P. (2008). Horribile dictum: Suomalaisen koulun menestystarina [The story of Finnish school success]. Didacta Varia, 13, 29–42.
79
Kontos, S., & Wilcox-Herzog, A. (1997). Teachers' interactions with children: Why are they so important? Young Children, 52, 4–12. Lerkkanen, M. -K., Poikkeus, A. -M., & Ketonen, R. (2006). ARMI – Luku- ja kirjoitustaidon arviointimateriaali 1. luokalle [ARMI – A tool for assessing reading and writing skills in Grade 1]. Helsinki: WSOY. Lyytinen, P., Eklund, P., & Lyytinen, H. (2005). Language development and literacy skills in late-talking toddlers with and without familial risk for dyslexia. Annals of Dyslexia, 55, 166–192. Muthén, B. O. (1991). Multilevel factor analysis of class and student achievement components. Journal of Educational Measurement, 28, 338–354. Muthén, L., & Muthén, B. O. (1998–2008). Mplus Version 5.1 & Mplus Users’ Quide. http://www.statmodel.com Myrberg, E. (2007). The effect of formal teacher education on reading achievement of 3rd-grade students in public and independent schools in Sweden. Educational Studies, 33, 145–162. National Board of Education (2008). Retrieved Dec. 15, 2008, from http://www.oph.fi/ english National Institute of Child Health and Human Development, Early Child Care Research Network (NICHD, ECCRN) (2002). The relation of global first grade classroom environment to structural classroom features and teacher and student behaviors. The Elementary School Journal, 102, 367–387. National Institute of Child Health and Human Development, Early Child Care Research Network (NICHD, ECCRN) (2004). Does class size in first grade relate to changes in child academic and social performance or observed classroom processes? Developmental Psychology, 40, 651–664. Nurmi, J. -E. (in press). Students' characteristics and teacher-child relationships in instruction: A meta-analysis, Educational Research Review. Nurmi, J., Viljaranta, J., Tolvanen, A., & Aunola, K. (2012). Teachers adapt their instruction according to students' academic performance. Educational Psychology, 32(5), 571. Retrieved from http://search.proquest.com/docview/1027413344?accountid=11774 Pakarinen, E., Lerkkanen, M. -K., Poikkeus, A. -M., Siekkinen, M., & Nurmi, J. -E. (2011). Kindergarten teachers adjust their teaching practices in accordance with children's academic pre-skills. Educational Psychology, 31, 37–53. Pianta, R. C. (2006). Classroom management and relationships between children and teachers: Implications for research and practice. In C. M. Evertson, & C. S. Weinstein (Eds.), Handbook of Classroom Management: Research, Practice, and Contemporary Issues (pp. 685–709). Mahwah, NJ: Lawrence Erlbaum Associates. Pianta, R. C., Howes, C., Burchinal, M., Bryant, D., Clifford, R., Early, D., et al. (2005). Features of pre-kindergarten programs, classrooms, and teachers: Do they predict observed classroom quality and child-teacher interactions? Applied Developmental Science, 9, 144–159. Poikkeus, A. -M., Ahonen, T., Lerkkanen, M. -K., Niemi, P., Siekkinen, M., & Nurmi, J. -E. (2006). First step study. University of Jyväskylä, University of Eastern Finland, and University of Turku. Pomerantz, E. M., & Murry Eaton, M. (2001). Maternal intrusive support in the academic context: Transactional socialization processes. Developmental Psychology, 37, 174–186. Pressley, M., Hogan, K., Wharton-McDonald, R., Mistretta, J., & Ettenberger, S. (1996). The challenges of instructional scaffolding: The challenges of instruction that supports student thinking. Learning Disabilities Research and Practice, 11, 138–146. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods. USA, California: Sage Publications. Rutter, M. (1997). Nature-nurture integration: The example of antisocial behavior. The American Psychologist, 52, 390–398. Rutter, M., & Maughan, B. (2002). School effectiveness findings 1979–2002. Journal of School Psychology, 40, 451–475. Scarr, S., & McCartney, K. (1983). How people make their own environments: A theory of genotype → environment effects. Child Development, 54, 424–435. Seymour, P. H., Aro, M., & Erskine, J. M. (2003). Foundation literacy acquisition in European orthographies. British Journal of Psychology, 94, 143–174. Skinner, E. A., & Belmont, M. J. (1993). Motivation in the classroom: Reciprocal effects of teacher behavior and student engagement across the school year. Journal of Educational Psychology, 85, 571–581. Snijders, T., & Bosker, R. (1999). Multilevel analysis. An introduction to basic and advanced multilevel modeling. Thousand Oakes, CA: Sage Publications. Stipek, D. J., & Byler, P. (2004). The early childhood classroom observation measure. Early Childhood Researh Quarterly, 19, 375–397. Sutton, R. E., & Wheatley, K. E. (2003). Teachers' emotions and teaching: A review of the literature and directions for future research. Educational Psychology Review, 15, 327–358. Tschannen-Moran, M., & Woolfolk Hoy, A. (2007). The differential antecedents of self-efficacy beliefs of novice and experienced teachers. Teaching and Teacher Education, 23, 944–956. Weinstein, R. S., & Middlestadt, S. E. (1979). Student perceptions of teacher interactions with male high and low achievers. Journal of Educational Psychology, 71, 421–431.