A professional learning model that cultivates primary science classrooms’ representational profiles

International Journal of Educational Research 76 (2016) 12–33

Contents lists available at ScienceDirect

International Journal of Educational Research journal homepage: www.elsevier.com/locate/ijedures

A professional learning model that cultivates primary science classrooms’ representational profiles Kim Nichols* , Robyn Gillies, Donna Kleiss The University of Queensland, School of Education, Queensland 4072, Australia

A R T I C L E I N F O

A B S T R A C T

Article history: Received 27 September 2015 Received in revised form 5 December 2015 Accepted 12 December 2015 Available online xxx

The aim of the study was to explore representational profiles of seven primary science classrooms prior to and following professional learning around representational practices. Teachers’ self-efficacy to teach year six geology with representations and their competencies to interpret, explain and choose geology-based representations significantly improved. Teachers with very different years of teaching experience chose the same types of representations to teach the same concepts. Despite variation between classes and students, there was a significant and substantial improvement in student competencies to interpret, understand and create representations to explain geological concepts and this improvement occurred across all classes. The findings suggest the professional learning promoted conceptual and representational competencies in all classes despite the differences across the teachers and their classrooms. ã 2015 Elsevier Ltd. All rights reserved.

Keywords: Representational competencies Representational profile Teacher professional learning Teacher self-efficacy

1. Introduction The communication of science requires a specialised language that is inextricably linked to several other modes of representation (diagrams, images, graphs, mathematical symbols) and the meaningful way they are used to construct knowledge or argument (Lemke, 2004) through inquiry. Often, modes of representation like graphs and diagrams are complex. Members of the scientific research community consistently work with, conduct scientific inquiry through, and communicate their discoveries by using relevant modes of representation. In other words, they are fully fluent with representational modes and capable of interpreting, explaining creating and relating the meaning of and across representations. However, teachers do not always have these skills and may not be able to explain representations to students or guide students to the key features of representations that convey conceptual meaning. Furthermore, students are often unable to interpret information from or work with representations in order to construct an understanding in science ([206_TD$IF] Jaipal, 2010). While some students possess the skills of reading particular or several representations in science and constructing meaning from them, many students require these skills to be modeled for them (Prain & Waldrip, 2006). To facilitate the teaching and learning of science concepts, teachers need to be explicitly trained to choose, sequence, and explain multiple content-appropriate and accepted representations of science including text, drawings, diagrams, graphs, tables, pictures and sound in animations to promote students’ interpretation, understanding, explanation of, and even creation of representations. Prain and Waldrip (2006) emphasize the need for teachers to “focus not only on the concepts,

* Corresponding author at: The University of Queensland, School of Education, St Lucia, Queensland 4072, Australia. E-mail addresses: [email protected] (K. Nichols), [email protected] (R. Gillies), [email protected] (D. Kleiss). http://dx.doi.org/10.1016/j.ijer.2015.12.002 0883-0355/ ã 2015 Elsevier Ltd. All rights reserved.

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

13

but on the signifying codes within different representations and their interrelationship” (p. 1864). Teachers’ inquiry-based instructional approaches in the science classroom must be structured around the introduction, discussion, and conceptual meaning of a range of representational modes; the explicit discussion of key representational features and related meaning between the accepted representations of the scientific community. 1.1. Instructional competencies around representations Developing students’ conceptual and representational competencies in science is an important instructional goal (diSessa, 2004) and one which requires skills in evaluating and understanding the semiotic and material affordances of accepted representations of science (Kozma & Russell, 2005). To be able to explain the features of a representation that convey meaning requires agency over representations (Kockelman, 2007). Previous research suggests that teachers and students must be trained to develop this agency (Nichols, Gillies, & Hedberg, 2015a; Nichols, Stevenson, Hedberg, & Gillies, 2015b). Agency with accepted representations of science is only a foundational skill upon which other important skills emerge. Once agency is attained, and this becomes an automatic tendency to evaluate accepted representations of science in terms of what they relay about a concept or do not relay, then careful choice, sequencing and conceptually connecting meaning across multiple representations is possible. In other words, agency precedes fluency with representations or the ability to translate meaning across representations to build a picture or knowledge of a science concept, idea or phenomenon. Several studies have shown the critical importance of instructional modeling of the language conventions around multiple representations and the concepts they convey for promoting students’ meaning making and conceptual understanding of science through representations (Airey & Linder, 2009;[207_TD$IF] Hilton & Nichols, 2011; Nichols, Hanan, & Ranasinghe, 2013a; Nichols, Ranasinghe, & Hanan, 2013b). However, there is a paucity of research around professional development or learning models that promote teachers’ representational agency and the consequent impact on teachers’ and students’ use of and competency to work with representations. Can professional learning help build these representational profiles of teachers and students? If so, what model of professional learning will support the development of classroom representational practices and profiles? Some more recent studies suggest that this may be possible. A study (Gillies, Nichols, & Khan, 2015) investigating the impact of professional learning of teachers around representational practices on year 6 students’ social and scientific language skills around geological concepts showed that students exhibited higher levels of social and scientific language than their untrained peers. A subsequent study (Nichols et al., 2015a) exploring how argumentation-promoting inquiry practices focused on representations impacted on year 6 students’ representational competencies and knowledge-building discourse around geological concepts showed that compared to their untrained peers, these students exhibited significantly higher conceptual understanding and collaborative knowledge construction discourse. Moreover, these students were better able to work with accepted representations to inquire about and construct an understanding of geological concepts. These skills, taken together, provided a measure of students’ representational fluency. 1.2. Professional learning for inquiry instruction It has been suggested (Crawford, 2012) that many professional learning models are not effective in supporting teachers’ understanding of the nature of scientific inquiry nor do they sufficiently prepare teachers to implement inquiry learning into their science classrooms, due to the focus on curriculum rather than pedagogical approaches. This claim suggests the need to consider what teachers require from professional learning experiences in order to design more effective models (Grigg, Kelly, Gamoran, & Borman, 2012). Several studies have researched what teachers perceive to be useful aspects of professional learning for improving their implementation of inquiry science. A study conducted by Tseng, Tuan, and Chin (2012) discussed interview responses from 15 experienced junior high school teachers to reveal their perspectives of and recommendations for professional learning around inquiry teaching. A seminal study by Supovitz and Turner (2000) described a meta-analysis of published research that identified widely accepted characteristics of quality professional learning. These studies assert that an immersion approach to professional learning is considered most effective. Such a professional learning model immerses teachers in inquiry learning themselves, demonstrates how learning links to specific curriculum standards, involves material and practical resources, demonstrations, and strategies that can be connected to other areas of learning, and provides sustained support from research teams. Supovitz and Turner (2000) analysed data collected through self-reported teacher surveys and found that increased engagement in professional learning which embodied these elements was associated with both increased use of inquiry-based teaching practices and higher levels of uptake of the professional learning. [208_TD$IF]Lee, Hart, Cuevas, and Enders (2004) implemented a professional learning intervention that embraced many of these elements in its design and found teachers identified positive changes to their practices in alignment with an inquiry approach to teaching and learning. The suggestion here is that a professional learning model that captures these successful elements around inquiry teaching with a focus on representational practices could be an effective model for improving representational and conceptual competencies in science.

14

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

1.3. Self efficacy around science instruction with representations This study focuses on professional learning of upper primary teachers as they face enormous challenges in teaching science using representations. These challenges are associated with the very high number of key learning areas they are responsible for teaching, their limited exposure to domain-specific content knowledge and language conventions in science (Osborne, Simon, & Collins, 2003), understanding of scientific inquiry (Bryan, 2003), low confidence which is documented through self-efficacy in teaching science and the limited time allocated to science in primary schools (Howitt, Rennie, Heard, & Yuncken, 2009). Self-efficacy generally involves ideas about how well one thinks he or she can accomplish tasks (Bandura, 1977). In the classroom, teacher self-efficacy affects choices individuals make, such as personal objectives and student learning goals (Tschannen-Moran & Hoy, 2001). A teacher’s self-efficacy level also influences teaching practice (McKinnon & Lamberts, 2014) and can affect choices or reasoning around representations and strategies used to engage students with representations. Therefore, it is critical for teachers to possess a good perception of their own efficacy (Schunk, 1991; Tschannen-Moran & Hoy, 2001), particularly around their representational practices. 1.4. Purpose of the study This study sought to explore how professional learning around representational practices in year six primary science classrooms could influence teachers’ and students’ representational profiles. For the purposes of the study, teachers’ profiles comprised representational competencies, practices around and reflections of using and choosing representations, and selfefficacy. Students’ profiles included representational and conceptual competencies, and demonstrated abilities to work with representations in small groups. In this study, a multiple case study inquiry of seven classrooms was conducted to explore and profile these competencies following professional learning around representational practices in order to assess the effectiveness of a professional learning model. The research questions for the study follow. 1. What elements of professional learning around representational inquiry science practices potentially cultivate classroom representational profiles? a What elements of professional learning cultivate teachers’ representational profiles? b What elements of professional learning improve students’ representational profiles?

2. Methodology This study employed a before and after convergent mixed methods approach (Cresswell, 2012). The study was carried out with seven Brisbane metropolitan primary science teachers from three schools having similar socio-demographic profiles. Quantitative and qualitative data were collected prior to and following a two day professional learning program as well as following the implementation of a science unit of work in term 3. Then, outcomes and analyses of the quantitative and qualitative data were merged to draw conclusions. Quantitative data included teacher and student representational and conceptual competency measures, efficacy surveys and coding of representational use in the classroom. Qualitative data included classroom observations and teacher stimulated recall interviews. Interviews were conducted following the implementation of the science unit. 2.1. Participants and procedure Participants included seven primary teachers teaching a year six unit about earth structure and geological events. Teacher participants included seven females and ranged in years of teaching experience from six months/less than one year to 20 years. The teacher participants included two beginning teachers (first year out of their degree), three mid-career teachers (between 8 and 10 years of teaching experience) and two senior teachers (with 15–20 years of teaching experience). Student participants included 67 girls and 55 boys with an age range from 10 to 12 years. Once ethical clearance was obtained for this study, participants were recruited based on several criteria. 1) The teachers were teaching year 6 in the year of the study. 2) The teachers provided written consent to participate in professional development and implement a unit on geological events. 3) The teachers and students consented to classroom observations, tests and interviews. The professional learning was conducted over a period of two days, and all teachers from the study cohort attended. At commencement of the professional development (time 1) teachers completed a test to determine their ability to evaluate, choose, sequence and explain representational strategies pertaining to the science unit to be taught in the following term. Teachers were also asked to complete a survey, which assessed their self-efficacy surrounding the teaching and resourcing of the geological science topic.

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

15

Across the course of the professional learning, teachers were led through a series of activities able to be readily implemented in their own classrooms. The teachers played the role of students conducting and participating in activities across a unit or work about earthquakes and associated geological events as this is deemed by learning theorists to promote successful development (Lieberman, 1995). In this way they were immersed in the inquiry science unit. The activities were structured around [209_TD$IF]Bybee’s (2013) 5E’s model of inquiry to facilitate their understanding of planning and implementing the geology and future science inquiry units. A series of activities aligned with this inquiry model were carried out across the two professional learning days and are described below. 2.1.1. See, scan, analyse The teachers experienced an engage activity called ‘see, scan, analyse’ designed to immerse them in the topic. They were given a picture of a newly built house that had a visible hole in its center with a hill in front of it that showed a visible dent in the ground on the top of the hill. In pairs, and with magnifying glasses, teachers were instructed to look over the picture and note what they could see, then use the magnifying glass to carefully scan the picture and analyze from their observations what had happened to the house. Pairs shared ideas with other pairs and then a whole class discussion engaged sharing of ideas to reach a common or shared understanding of the mystery. 2.1.2. Pick the science A second engage activity was then initiated called ‘pick the science’ where teachers were asked to pick science terms or descriptions from a museum-like display of commentaries authored by New Zealand year 6 students that had experienced the 2010 earthquake. Words like earthquake, magnitude, mud volcanos, liquefaction and shaking earth were shared by teachers and included on a word wall. Teachers were asked to write these words in their science journals in their glossary and then asked to share what they thought they knew or wanted to know about earthquakes. This was recorded in a large graphic organiser on display. 2.1.3. Modeling earth structure and plate boundaries Teachers in pairs ‘explored’ a series of interactive digital learning objects on the Earth’s structure and the various types of plate boundaries. They were asked to answer questions and draw and label the earth. They were instructed to draw the different plate boundaries, then construct them with plasticine and labels so they could ‘explain’ what was happening at the boundaries using the models and scientific language. 2.1.4. Evaluating representations In order to ‘elaborate’ their understanding, teachers were presented with multiple different representations of the earth structure as cut-away diagrams and asked to ‘evaluate’ them in terms of their form, features, potential benefits and drawbacks in terms of what they conveyed or didn’t convey. Teachers were asked to choose what they perceived as the best representations and explain why and how they would use them for teaching and learning (e.g. what pedagogical strategies they would employ around the representations). This evaluation activity was repeated using multiple different representations of different types of plate boundaries. 2.1.5. What are earthquakes? Teachers then explored the nature and definition of earthquake through an online interactive animation (demonstrated from the front of the classroom) showing a cut-away section of land. At a click an earthquake starts at the ‘epicenter’ and ‘waves’ radiate out in all directions through the earth causing ‘shaking’ of the earth which is visible and audible in the animation. Teachers were asked to use the animation and associated text to write in their journal a definition of earthquake with a diagram. A whole class discussion considered an appropriate definition of earthquake. 2.1.6. How are earthquakes measured? A worksheet was used to walk teachers through how earthquakes are measured featuring a ‘Modified Mercali scale’ interactive animation and description of Mercali the scientist. The animation showed a street of buildings with a Roman numeral scale below from 1 through 12. Selection of these scale levels showed different levels of intensity of an earthquake or the extent of damage it would cause to the structures in the animation. Next the worksheet was used to step teachers through an online interactive animation showing a street of residential houses. An interactive scale from 1 through 10 could be selected that demonstrated magnitude or strength of an earthquake and that showed relative damage to the houses. The worksheet then introduced the Richter scale and the scientist who invented the scale. Teachers were asked to work in pairs to define magnitude and intensity of an earthquake giving examples and to record it in their journals followed by a whole class discussion. Teachers were asked if a high magnitude earthquake occurred in an area without infrastructure (not built up) what the intensity of the earthquake would be. 2.1.7. Where do earthquakes occur? Teachers were shown the United States Geological Survey (USGS) site that highlighted where earthquakes occur across the world within the past week, days or hours. A worksheet stepped them through the site and had them compare earthquakes in Australia and surrounding countries in relative terms of numbers and magnitude.

16

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

2.1.8. Modeling earthquake activity Teachers participated in creating a model of a seismogram that records earthquake activity using a movable table, paper and pen (to record Table shaking). They were instructed to make sense of the seismogram in pairs about what the trace means and decide on appropriate ways of representing what happened. They were asked to think about the results on the trace and reach a shared understanding of what it means in terms of the relationship between the simulated ground (table) shaking and the nature of the ‘seismic-like’ waves on the trace (weak and strong ground shaking was simulated when creating the trace). Teachers were instructed to divide the trace into sections and explore or represent the progress of the earthquake. They were asked to represent (e.g. draw, label, describe, measure) any differences in the waves in each section and use these different representations as evidence to support their argument about how the nature of seismic waves relate to different strengths of ground shaking. Teachers measured the waves in the different sections then represented the results using a graphical representation. A whole class discussion explored the nature of the data and the appropriate way to represent it in a Table and graph. 2.1.9. Tsunami and liquefaction Teachers watched a BBC production of ‘How the 2011 Japan earthquake happened’. This was a good consolidation of what was already learned about plate boundaries, earthquake and magnitude definitions. But it also introduced associated geological events such as liquefaction and tsunami. Following on from the BBC video, teachers were shown a few animations on tsunamis and asked to evaluate which they thought were the most appropriate for their students and why. If they chose more than one animation, they were asked to decide how they would sequence them in a learning progression and what pedagogical strategies they would employ around them. Teachers then explored liquefaction through a hands-on modeling activity that uses a plastic cup, sand, a pie plate and water. The teachers modeled liquefaction and engaged in a ‘predict, reason, observe and explain’ strategy. Teachers were shown some images to help visualize what happens during the process of liquefaction at a sand grain level and then asked to explain the process. A video of a scientist talking about liquefaction as a process and why it is important to research it was shown and discussed. In addition to the activities the teachers participated in across the two days, researchers were consistently available to teachers to discuss any queries or concerns they had about the unit activities, or clarification they needed around activities as the unit proceeded. Visits to collect data across the unit were also opportunities to debrief and check in on the teachers’ needs or concerns. The intent was for teachers to have a sense that they could consult with the researchers at any time and that there was ongoing contact between researchers and teacher participants. At the completion of the professional learning (time 2) teachers were asked to complete the same test and survey delivered at time 1. This was again repeated at the conclusion of science unit (time 3; approximately four months after the professional learning days). Two lessons were video recorded across the unit in each classroom. The filmed lessons were transcribed and each was coded according to the various types, or modes of representations utilised within the space of a single lesson. Stimulated recall interviews were carried out by researchers at the conclusion of the study period to ascertain teachers’ perceptions of using and choosing representations to demonstrate key scientific ideas of the unit. Students were given a test to determine their conceptual and representational competencies (coded for interpreting, explaining and creating representations) at the start and at the end of the unit. 2.2. Instruments The Science Teaching Efficacy Beliefs Instrument (STEBI) was adapted from the original survey created by Riggs and Enochs (1990) for primary teachers. The STEBI is a reliable and valid instrument for measuring primary teachers’ efficacy toward teaching science concepts (Posnanski, 2002). The original STEBI includes 25 items. The instrument utilized in this study included 12 Likert-scale items (see Appendix A) comprising five response categories ranging from strongly agree to strongly disagree. In order to be consistent with the original STEBI items, four of the included items were written in a positive stance and eight of the included items were written negatively. In order to focus more precisely on the unit science content the teachers were teaching, they were told to only consider the “science” related to the unit of work they were teaching. The teacher representational competencies test included a series of questions designed to explore teachers’ conceptual understanding of and ability to make conceptual connections between representations. Appendix B provides an outline of the test provided to the teachers. The questions of the teacher representational competencies test evaluated the teachers abilities to interpret, compare, choose, sequence and explain the choices and strategies they would utilize around accepted representations and to relate meaning across different representations. The test was validated as a framework for teacher representational competencies in a previous study (Nichols et al., 2015[210_TD$IF]b ). To assess students’ conceptual and representational competencies, they were asked to complete a test prior to the start of and at the completion of the unit. The test is provided in Appendix C and was utilised by the teachers as assessment for the unit. The test was developed and validated in a previous study (Nichols et al., 2015a). Within the test, different items were designed to explore different representational competencies including interpreting representations (items 3–7) where students were asked to simply label diagrams, or match words to pictures; explaining representations (items 4–7) where students were required to explain the science knowledge that a given representation was conveying; and creating representations (items 1, 2 and 5) where students were asked to construct a representation to explain concepts. Each of these representational competency variables was scored using a combination of specific test items. The test score for each

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

17

representational competency (interpreting [16], explaining [18] and creating representations [16]) divided by the total possible score across all items for that competency was used for comparison pre- to post-unit. Video clips of each teacher using representations to teach in the classroom were shown to individual teachers in order to stimulate recall of their reasoning and pedagogical intentions with representations in the classroom. 2.3. Data analysis Teachers completed the self-efficacy survey at time 1 (at the start of the professional learning), time 2 (immediately following the professional learning days) and time 3 (immediately following implementation of the unit). Responses were analyzed for individual questions. As the sample cohort of teachers was small, the non-parametric Friedman’s test was used to investigate differences across the three time points. For pair-wise comparisons between any two time points, a Wilcoxon signed rank test was conducted with a Bonferonni adjusted alpha level of 0.017. The teacher representational competencies were analysed using the Structure of Observed Learning Outcomes (SOLO) taxonomy (Biggs & Collis, 1982) in order to assess the level of conceptual understanding of and ability to make conceptual connections between representations prior to and following the delivery of the unit. The test responses were analysed iteratively. They were coded individually by two senior science educators and followed by a discussion to reach a consensus. The SOLO-taxonomy describes levels of complexity in an individual’s conceptual understanding of a subject, through five levels (Table 1), and it is claimed to be applicable to any subject area. This analytical process of coding conceptual depth through interpretations of representations using the SOLO taxonomy was based on a validated schedule that was developed by Rundgren, Hirsch, Chang Rundgren, and Tibell (2012) and modified by Nichols et al. [21_TD$IF](2015b). Given the small sample size and discrete nature of the coding, median scores were compared across the three time points using the non parametric Friedman test. Comparative ranks between time points were conducted using a Wilcoxon Signed Rank Test. Each question on the student representational competencies test was provided a score based on the level of complexity of the answer required. For conceptual understanding, the entire test score divided by the total possible score for the entire test was used for comparison between the time points. Each of the representational competency variables (interpreting, explaining and creating representations) was scored using a combination of specific test items. The data for the students’ conceptual and representational competencies collected prior to (time 1) and at the conclusion of (time 2) the science unit possessed a hierarchical structure where students were nested within classes. Therefore, multilevel analyses (Goldstein, 2005) were undertaken to compare differences across the two time points and to determine the degree of students’ improvement on performance measures across classes. Video recorded lessons were coded for various phases as well as for frequency of representation types utilised across the lessons. The average lesson ran for 55 min and so there was approximately 110 min of lesson time recorded for each teacher. The frequency of representation category used in each of the lesson phases was also explored for particular patterns of use. Table 2 shows the various phases that were evident across video recorded lessons and the description of each phase. Table 3 shows the coding scheme used for representation categories and the resources utilised by the teachers in their lessons. The stimulated recall teacher interview responses were analysed with segmenting, coding and the development of category systems in QSR NVivo. Facesheet codes were utilised in classifying themes as they emerged. In NVivo, nodes were constructed to show the nature of the emergent themes that reflected teachers’ perceptions of using and choosing representations in their classrooms.

Table 1 SOLO levels of conceptual understanding around representations. SOLO Level

Associated conceptual understanding

Level 1 Pre-structural

Individuals are simply acquiring bits of unconnected information, which have no organization and make no sense. These individuals are able to draw ideas and information from separate representations but are unable to make any conceptual connections between any two representations Simple and obvious connections are made, but their significance is not grasped. Individuals at this level can make simple connections between representations that relate more to structure rather than conceptual ideas A number of connections may be made, but the meta-connections between them are missing, as is their significance for the whole study object. This individual is able to make structural and conceptual connections between two representations but not beyond that across multiple representations The individual is now able to appreciate the significance of the parts in relation to the whole. An individual at this level is able to relate the ideas across multiple representations within the topic but not extend them across subject domains The individual makes connections not only within the given subject area, but also beyond it, and is able to generalize and transfer the principles and ideas underlying the specific instance. This individual can relate and extend conceptual ideas across representations within the topic of the unit and beyond to other topics they are teaching

Level 2 Uni-structural Level 3 Multistructural Level 4 Relational Level 5 Extended Abstract

Adapted from Biggs and Collis (1982).

18

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

Table 2 Lesson phase coding scheme. Lesson phase

Description

Lesson intent

Teacher introduces what is to be taught during the lesson

Revision

Revision of previous lesson Revision of prior concepts taught during the unit Summary of what has been taught prior to the current lesson

Engage

An activity designed specifically to engage students in the learning process Generally involves student participation (not necessarily the whole class) Generally includes use of a representational model (visual or physical)

Introduction

The new science concept is introduced to the class Generally overtly taught by the teacher May include representational discourse, such as visual texts

Explanation

The teacher explains the science concept in more detail Includes teacher-directed explanations of a specific concept, using representational models, including diagrammatic representations, purposeful gestural movements, visual texts etc Often includes teacher-directed discussion with the class, including teacher-student interaction

Consolidation (student-centred activities)

Students consolidate or practice the newly introduced science concept This phase may include any of the following: Students construct text either independently of in a small group Students work independently or in a small group to research an aspect of the topic just introduced Students construct a physical model representing the introduced science concept

Elaboration

The teacher elaborates on the newly introduced scientific concept He/She may expand on the ideas in the explanation phase, or may explain one aspect in more detail This phase generally occurs immediately after the explanation phase or after the consolidation phase This phase may occur in response to the immediate needs of the students and so, may not be planned Teachers refer to representational models in this phase to assist students to understand the new scientific concepts

Review of lesson

The teacher reviews the lesson, highlighting the key points or scientific concepts necessary for student understanding This phase is generally teacher-directed, with little or no student interaction

Reflection

The teacher asks students to reflect on their understanding of the newly introduced scientific concept The teacher will generally ask the class to use gestural movements to indicate their level of understanding with the main content/concepts of the lesson

Note: these ‘lesson phases’ were identified based on observations of lessons across the study cohort and are not to be confused with the ‘5E’s phases across a unit of work’.

3. Results 3.1. Teacher self-efficacy Questions 2, 7, 8, 9, 10 and 12 were significantly different across the three time points (time 1 [T1], time 2 [T2], and time 3 [T3]). Results are summarized in Table 4. With pair-wise comparisons, only Question 10 showed significant differences in representational competencies scores between T1 and T3, z = 1.414, p < 0.015 with a large effect size (r = 0.71). There were no significant differences between T1 and T2 z = 1.00, p = 0. 317, r = 0.5 or T2 and T3, z = 1.341, p = 0.18, r = 0.671. The teachers disagreed more strongly (p < 0.004) with the item “I would not invite the Principal to evaluate my instruction” at the end of the study indicating greater confidence to have the Principal view their instruction of the unit of work they were trained to implement. Next to item 10, item 9 “I am wondering if I have the necessary skills to teach science” is strongly significant (p < 0.022) across the three time points which indicates that they wonder less about their ability to be effective in their science instruction over the time of the study period. Item 2 is also strongly significant over the three time points (but not across any two time points) (p < 0.015) “I know the steps necessary to teach science concepts effectively”. This is a curious outcome but could indicate that they perceived they had much to learn about the sequencing of content or ideas when teaching science concepts. 3.2. Teacher representational competencies Teacher representational competency test responses at Time 1 (T1), Time 2 (T2), and Time 3 (T3) were compared using the Friedman test. A Wilcoxon Signed Rank Test was performed for Post-hoc comparisons with a Bonferonni adjusted alpha level of 0.017 (0.05/3). There was a statistically significant difference in the test’s representational competencies scores across the

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

19

Table 3 Coding scheme for representation categories and examples of the resources utilized. Type of representation

Resources/tools

Concrete representation Representation with concrete models, such as hands-on models and concrete physical models. 3D Plasticine/play doh representations that can be held and touched Physical models of the Earth Ball and stick models Pictorial representation [194_TD$IF]Use of visual systems of meaning to represent or support representation of concepts related to science, such as graphs, diagrams, 2D pictorial representation, animations, moving footage

Mathematical/symbolic representation [194_TD$IF]Use of mathematical or scientific notation systems to represent or support representation of concepts related to science, such as equations, notations and variables without a detailed description Realistic representation Representation through realistic, real-world, experienced contexts or metaphors

Photographs, Diagrams Graphs Maps, Charts, Tables Graphic organisers Animations Moving footage

Numerals Algebraic symbols Chemical notation

Narrative verbal connection between an experience with a science concept Verbal connection between a tangible, known thing

Gestural/[195_TD$IF]kinaesthetic representation Purposeful use of hand or body movements to represent or support the representation of concepts Purposeful hand movements related to science Purposeful body movements Configuration of body to model a specific concept

Table 4 Time 1, Time 2 and Time 3 median (inter-quartile range) scores for the STEBI survey (N = 7). Time 1

Time 2

Time 3

x2

p

4.00 (3.00– 4.00) Q2.I know the steps necessary to teach science concepts effectively [196_TD$IF]3.00 (3.00– 4.00) Q3.I am continually finding better ways to teach science [197_TD$IF]4.00 (3.00– 5.00) Q4. I am not very effective in monitoring science experiments [198_TD$IF]4.00 (3.00– 4.00) Q5.I generally teach science ineffectively 4.00 (4.00– 5.00) Q6. I understand science concepts well enough to be effective in teaching elementary science 4.00 (4.00– 4.00) Q7.I find it difficult to explain to students why science experiments work 4.00 (3.00– 4.00) Q8.I am typically able to answer students’ science questions 4.00 (3.00– 4.00) Q9.I wonder if I have the necessary skills to teach science 4.00 (3.00– 4.00) Q10.Given a choice, I wouldnotinvite the principal to evaluate my science instruction 3.00 (2.00– 4.00) Q11.When a student has difficulty understanding a science concept, I am usually at a loss as to 4.00 how to help the student understand it better (3.00– 4.00) Q12.When teaching science, I usually welcome student questions 4.00 (4.00– 4.00)

4.00 (4.00– 4.00) 4.00 (4.00– 4.00) 4.00 (4.00– 5.00) 4.00 (4.00– 4.00) 4.00 (4.00– 5.00) 4.00 (4.00– 4.00) 4.00 (4.00– 4.00) 4.00 (4.00– 4.00) 4.00 (4.00– 4.00) 4.00 (4.00– 4.00) 4.00 (4.00– 4.00) 4.00 (4.00– 5.00)

4.00 (4.00– 4.00) 4.00 (4.00– 5.00) 5.00 (4.00– 5.00) 4.00 (4.00– 4.00) 4.00 (4.00– 5.00) 4.00 (4.00– 4.00) 4.00 (4.00– 5.00) 4.00 (4.00– 4.00) 4.00 (4.00– 5.00) 4.00 (4.00– 5.00) 4.00 (4.00– 4.00) 5.00 (4.00– 5.00)

4.67

0.097

8.38

0.015

5.69

0.058

3.80

0.150

0.50

0.779

0.67

0.717

6.00

0.050

6.00

0.050

7.60

0.022

Survey question Q1. Even when I try very hard, I don’t teach science as well as I do most other topics

Note: x2 = Chi square. 1 = strongly agree, 2 = agree, 3 = undecided, 4 = disagree, 5 = strongly disagree. Bolded text denotes significance.

11.27 0.004

4.67

0.097

6.00

0.050

20

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

three time points, x2 (2, n = 7) = 11.565, p < 0.003. Inspection of the median values showed an increase in competencies scores from T1 (Median = 2) to T2 (Median = 3) and an increase at T3 (Median = 4). A Wilcoxon Signed Rank Test revealed a statistically significant difference in representational competencies between T1 and T3, z = 2.460, p < 0.014 with a large effect size (r = 0.66). There were no significant differences between T1 and T2 z = 2.121, p < 0.034, r = 0.57or T2 and T3, z = 1.857, p = 0.063, r = 0.50. 3.3. Lesson distribution of representation categories Fig. [21_TD$IF]1 shows the representational use profile for teacher 1 from school 1 (T1/S1) and teacher 2 from school 2 (T2/S2). These two teachers taught the same two lessons in The Changing Earth unit. However, their approach to the lessons was unique. Lesson 1 introduced the different types of plate boundaries (divergent, transform and convergent) using a Scootle digital learning object. Each teacher showed the learning object using the interactive whiteboard in the front of the classroom. T2 had her students work in small groups to construct a divergent boundary using play doh and then to explain what happens at that boundary using their model and record it in their science journals. T1 had students write a definition of divergent boundary in their science journals. She showed students a tectonic plates map of the world and had students work individually to trace and label the divergent boundaries on the map. She then discussed and named some of those

[(Fig._1)TD$IG]

Fig. 1. Profiles of representation category use for teacher 1 in school 1 (T1/S1) and for teacher 2 in school 2 (T2/S2). The x-axis is a direct count.

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

21

boundaries. Students then worked on computers with the digital learning object and filled out a worksheet. Both T1 and T2 reviewed with the students the main messages of the lesson. In lesson 2 students learned about liquefaction by doing a practical activity. The graphs in Fig. 1[213_TD$IF] suggest that T1 utilized more representations to teach her lessons than T2. T1 is in her first year of teaching, T2 is a mid-career teacher. T1 was more convinced of the potential of representations to support learning and engagement of her students. T2 felt she had become more discriminating about the representations she chose for her lessons. T2 had indicated that her lessons were structured identically in order to manage behavior in her classroom. Her lessons consisted mainly of a revision of the last lesson, an introduction with explanation, consolidation using a student hands-on activity and then review. This lesson structure would make it difficult for her to utilize more representations as the representations would need to fit within the parameters and intent of the lesson phases that she sustains. T1 had not yet moved into a consistent lesson phase pattern. Both T1 and T2 have a similar profile in terms of the representation categories that were utilized to teach the lesson. This suggests that the types of representations are related to the content or concept being taught and that the professional learning effectively imparted this critical notion to the teachers in the study. The representation use profiles of teachers across the study teaching the same content followed this pattern and show that teachers were aware that different scientific concepts have different but unique multi-media literacy demands and this is reflected by the similar choice of representation categories across teachers teaching the same concepts. The next level of analysis was to explore the average number of representations (all representation categories) used across the lesson phase categories. Fig. 2[214_TD$IF] shows all of the lesson phases that were observed across all the recorded lessons in both units. The order of the lesson phase categories on the x-axis reflects the order or sequence in which these phases typically occurred. The lesson phase categories that review previous or current lesson content, clarify the intent of the current lesson or reflect on the lesson content had the fewest mean total representation use. These phases typically occur at the outset and end of a lesson. Teacher explanations or consolidation activities had the highest mean total representation use with explanations using an average of 14 representations for that phase and consolidation activities using an average of 9 representations. These occurred in the main body of the lesson. In between these two extremes in average use of representations are the phases that lead up to or wrap up from the explanation and consolidation phases, engaging students into the topic, introducing the topic, and elaborating on the content or concept being taught. 3.4. Teachers’ reflections of using representations in the classroom Teacher stimulated recall interview responses (see Table 5) revealed their reflections of using representations in the science classroom. The main themes or categories of responses that emerged could be classified as Reasoning around the use and benefits,Criteria for choosing and Criteria for sequencing. Table 5 shows these response categories, some of the more typical responses along with the teacher and school where the response arose. These responses indicated that teachers were thinking about their representational practices, reasoning about the benefits and importance of careful selection and sequencing of representations, and the types of representations they noticed students enjoyed more or that made learning science more accessible. These responses are consistent with the significantly and substantially improved representational competency test scores. 3.5. Student conceptual and representational competencies Likelihood ratio tests revealed that (1) there was a significant variation between the classes in the study on student measures of conceptual understanding (CU), x2 (1) = 35.18, p < 0.001, interpreting representations (IR), x2(1) = 16.78,

[(Fig._2)TD$IG]

Fig. 2. Mean number of representations used in each lesson phase category. Y-axis is a direct count.

22

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

Table 5 Coding categories and teacher responses with the corresponding teacher and school code. Coding category

Responses

Teacher code

Reasoning and benefits

When I would present a diagram or representation to the kids, if they didn’t sort of understand that straight T3 away we’ll sort of, you know, have a bit more focus on that break down on what is going on in that representation Every child learns differently, so they might pick something up from one model and then at a different T2 representation they’ll pick something else entirely different You have to think really carefully before you show kids [a representation] because it is just as easy to confuse T1 them as it is to help them understand, so you have to be careful

School code S2

S3 S1

Choosing

Things that are going to grab the kid’s attention straight away Content appropriate for this age and to cater to different ability levels If I can get more out of one representation then I will, that’[19_TD$IF]s the one I would go with I like to use static diagrams to show areas for labelling and technical vocabulary, models to show more realistic 3D effects, simulations to demonstrate science phenomena in action, photographs to show real life images and film clips to demonstrate active phenomena

T5 T4 T7 T6

S3 S3 S3 S3

Sequencing

One that is content specific and also explains clearly and is something that can be built on or used to T4 demonstrate understanding. One that stimulates discussion I thought that the second one actually showed it more scientifically . . . and that the kids would be able to sort T2 of get a better understanding of what was actually happening between the particles I like to use basic diagrams to begin, with provision for elaboration and building of vocabulary or concepts. Use T6 of color to differentiate areas in 2D, ideally with progression to 3D. Realistic images and films. Simulations to demonstrate large scale concepts at a micro level

S3 S2 S3

p < 0.001, explaining representations (ER), x2 (1) = 28.87, p < 0.001 and creating representations (CR), x2 (1) = 17.68, p < 0.001; and (2) there was no significant difference between classes, from time 1 to time 2, in the mean increase of IR x2 (2) = 2.79, p = 0.25 and CR x2 (2) = 1.28, p = 0.53. This means that (1) a multilevel regression model was necessary to analyze the nested nature of the data (students nested in classes) and (2) a random intercept model (Rabe-Hesketh & Skrondal, 2005) was appropriate for analysis. To further justify the use of a multilevel model, non-parametric Shapiro–Wilk tests were conducted to assess normality of the random effects. In all cases, a non-significant result was found indicating normality. A random intercept model was used to explore how the professional learning and implementation of the inquiry unit may have contributed to students’ representational and conceptual competency measures at time 2 (at the end of the unit). To control for a student’s baseline measures, each of the time 1 (at the start of the unit) representational and conceptual competency measures were included as an independent variable. Table 6 shows that upon examination of students’ conceptual and representational competency measures, there were significant differences between time 1 and 2, with significant improvement on all measures. As shown in Table 6, there was an estimated standard deviation of 7.6 among average class CU scores about the initial grand average score of 20.3. Despite this variance, the restricted maximum likelihood (REML) estimation of the random intercept model revealed that there was a significant and substantial (approximately double) increase in the CU measure at time 2 with a large effect size (r = 0.74). This is also true for all representational competency measures each with large effect sizes for IR (r = 0.70), ER (r = 0.57) and CR (r = 0.72), and where the REML estimations revealed a significant and substantial increase in scores at time 2 (see Table 6). 3.6. Student work with representations Students in all classes were asked to work in groups on assigned inquiry-based tasks. In one activity the students were asked to analyze a fictitious trace of an earthquake that they created themselves by shaking a Table and recording the movement. Students were asked to make meaning of the trace relating it to scales that measure earthquake characteristics and using mathematical reasoning. The following vignette is an excerpt of a group questioning what the ‘ups and downs’ on the trace mean and then coming to a group consensus based on visual modeling (both with what was supplied and also with some newly created representations). Each student (S) in the group is designated by a number. S1: [Referring to the fictitious earthquake trace] What does it mean when it goes up and down and up and down? (S. interpreting the features of the representation/trace) S2: And up and down. S3: It means it’s an earthquake! (S. explanation of the representation) S1: Oh yeah . . . you’re not explaining this. What happens? . . . Ok, what happens if it goes like that? [S. draws/creates another representation to illustrate his question on a scrap piece of paper] (S. creates a representation to explain question) S3: An earthquake happened. (S. explanation of the representation) S4: It goes like that . . . Aw it means like . . . you know how it depends on . . . it’s going like this [S. uses a pen to draw another line to illustrate her point] (S. constructs visual representation to analogize and explain phenomena) S1: Nope . . . So what’s it going up and down mean? What’s the up and down . . . (S. persists in seeking explanation of this feature of the representation)

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

23

Table 6 Restricted maximum likelihood estimates of effects of the intervention on student scores for conceptual understanding, interpreting representations, explaining representations and creating representations under the random intercept model. [20_TD$IF]Conceptual understanding Fixed effects

Coefficient (SE)

95% CI

p-Value

Intercept Time

20.3 (3.2) 26.0 (1.9)

13.7–26.8 22.2–9.7

0.0003 <0.0001

Random effects

Std. dev.

95% CI

Between classes Within classes

7.6 13.6

4.0–13.8 12.3–15.0

[201_TD$IF]Interpreting representations [20_TD$IF]Fixed effects

Coefficient (SE)

95% CI

p-Value

Intercept Time

19.5 (3.2) 31.3 (2.4)

12.8–26.1 26.5–36.1

0.0003 <0.0001

Random effects

Std. dev.

95% CI

Between classes Within classes

7.2 17.4

3.3–13.4 15.7–19.2

[203_TD$IF]Explaining representations [20_TD$IF]Fixed effects

Coefficient (SE)

95% CI

p-Value

Intercept Time

16.1 (3.5) 15.6 (2.3)

8.8–23.2 11.1–20.1

0.002 <0.0001

Random effects

Std. dev.

95% CI

Between classes Within classes

8.2 16.3

4.2–14.9 14.8–18.0

[204_TD$IF]Creating representations [20_TD$IF]Fixed effects

Coefficient (SE)

95% CI

p-Value

Intercept Time

28.3 (3.1) 31.0 (2.3)

21.9–34.5 26.6–35.5

<0.0001 <0.0001

Random effects

Std. dev.

95% CI

Between classes Within classes

6.9 16.3

3.2–12.8 14.8–18.0

CI = confidence interval, SE = standard error.

S4: It’s an earthquake . . . (S. explanation of the representation) S1: The further it goes up, the further it goes down . . . (S. interprets the representation) S3: The more it shakes . . . S1: Yes exactly. (S. affirms) S2: It’s the magnitude. (S. explanation of the observations on the trace) The student interactions across this vignette consists of 3 interpretations of a representation (trace), 4 explanations and 2 instances of creating a representation to reason or explain the scientific phenomenon under investigation; clear demonstrations of representational competencies. As part of the social interactions and discourse in this group activity, the students are quite at ease with creating or constructing different representations to explain the science phenomena. Their conceptual understanding of earthquake magnitude is also clear. It is evident from these vignettes and the student representational competencies test outcomes that representational and conceptual competencies are mutually reinforcing skills and students can effectively work or reason with representations to make meaning. 4. Discussion and implications The professional learning provided to the teachers in this study was designed to improve core teacher representational competencies (Nichols et al., 2015a,[215_TD$IF]b) around accepted scientific representations associated with earth structure and geological events. These core competencies include evaluating representations, choosing representations, sequencing, explaining and translating meaning across different representations, as well as integration of representations and strategies into classroom teaching of science content. Teachers in this study achieved varying levels of conceptual understanding of

24

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

these core competencies, and across the three time points, the median level achieved was significantly different. The results of the teacher representational competencies test revealed that competencies such as evaluating, choosing, sequencing and explaining representations (or agency with representations) and translating meaning across representations to explain science concepts (or fluency with representations) improved significantly over the study period (with a large effect size). The greatest average number of representations utilised in a lesson appeared to be when the teachers were explaining science concepts to students or when there were consolidation activities being carried out. This is consistent with previous studies that have shown that the greatest use of representations is associated with explicit teaching (Jaipal, 2010). A study by Nichols et al. [21_TD$IF](2015b) around the nature of the development of competency to fluency with accepted representations of science in teachers, described a general pattern or series of stages. 1. Novice. Uses and creates representations for learning science but isn’t aware of his/her practices around representations (i.e. does not consider how he/she is making meaning using representations in the classroom). 2. Experienced beginner. Has an awareness of the importance of representations to learning science and is working on developing his/her practices (i.e. realises the semiotic potential of representations and the need to develop better practices). 3. Practitioner. Has awareness of the material and semiotic affordances of representations and is working on not only his/ her practices around them but his/her reasoning and explanation building using representations in the classroom. 4. Knowledgeable practitioner. Not only able to reason and explain with representations and is knowledgeable of their material/semiotic affordances but is working on developing students’ skills. 5. Expert. Has representational fluency. Is able to make conceptual links and connections between multiple representations, select, sequence and explain multiple representations and is working on developing students’ representational fluency. On the developmental scale for competencies with representations, the level of representational competency generally reached by teachers in the study was that of ‘Knowledgeable practitioner’. At this level the teacher is not only able to reason and explain with representations and is knowledgeable of their material/semiotic affordances but is working on developing students’ skills. The representational competency test results, interview responses and practices in the classroom confirm these levels had been achieved. Training teachers to foster the development of representational competencies (evident in the representational competencies test scores) was associated with a positive perception around the use of representational practices following professional learning. Stimulated recall interview responses confirmed the test results and revealed that representational reasoning and epistemologies were enhanced. This is consistent with previous findings (McLean & Rowsell, 2013) demonstrating purposeful training in modes and semiotic affordances of representations within pre-service primary teacher education courses promotes knowledge around representational practices and design principles. The shift in reasoning about the choice, sequencing and explanation of representations evident in the interview responses in this study is critical if teachers are to change the way they construct knowledge in the science classroom (Hubber, Tytler, & Haslam, 2010). Students need conceptual and representational competencies (diSessa, 2004) to learn science effectively but they will only attain this if teachers are able to induct them into the representational conventions and practices in science. The analysis across the three time points of the teachers’ responses to the self-efficacy survey items indicated that there was a significant shift on several items and suggests an increase in confidence to teach science using representations. The analysis between time points indicated that this shift was mainly reflected in the item that prompted teachers to consider having the Principal observe their lessons across the unit they were teaching. There was an increasing confidence for this to occur. Teachers became progressively more negative about the item “I know the steps necessary to teach science concepts effectively”. This could indicate that having a deeper appreciation of the science content they were teaching and the representational practices needed to teach it, they still felt they had much to learn about the sequencing of content or ideas when teaching science concepts. The findings indicated that teachers were able to successfully foster representational and conceptual competencies in their students in the four month period following professional learning. Students’ tested conceptual understanding, interpreting representations, explaining representations and creating representations measures significantly and substantially improved (with large effect sizes) across the unit. These competencies were also evident in student discourse during small group activities where students were able to interpret and explain representations provided to them, and engage in meaningful discussion and exchange of ideas through constructed representations. Despite the very significant variation in all students’ representational and conceptual competency measures between the classes and the known differences between the teachers in terms of their years of teaching experience, there was a substantive mean improvement on all student measures from time 1 to time 2 across all classes. This suggests that the professional learning intervention was successful in improving students’ representational profiles across all classes. The representational use profiles of teachers teaching the same lesson content were similar, their reflections of and reasoning around the use of representations in the recall interviews demonstrated good representational competencies and their representational competency measures significantly improved (with a large effect size); together suggesting the professional learning influenced the teachers’ representational practices and profile. The professional learning model described in this study included successful elements of professional development for inquiry teaching such as encouraging the teachers to involve themselves in the inquiry as they would involve their students;

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

Appendix A.

25

26

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

27

28

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

29

30

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

31

32

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

providing a rich source of practical resources, demonstrations and strategies that connect to the curriculum standards and maintaining support over the period of the study. Embedded in the professional learning were opportunities to inquire through and work with representations as their students would do, to explain science using representations (both accepted and created), to evaluate accepted representations associated with the unit’s science concepts and to articulate their learning and ideas of pedagogical strategies around representations. These embedded opportunities to build representational and conceptual competencies may have significantly improved their representational profiles and those of their students. Lieberman’s (1995) profound statement about transforming conceptions of professional learning resounds strongly with the findings of this study; “the way teachers learn may be more like the ways students learn than we have previously understood” (p. 68). Acknowledgement This study was funded by the Queensland Department of Education and Training. References Airey, J., & Linder, C. (2009). A disciplinary discourse perspective on university science learning: achieving fluency in a critical constellation of modes. Journal of Research in Science Teaching, 46(1), 27–49. Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioural change. Psychological Review, 84, 191–215.

K. Nichols et al. / International Journal of Educational Research 76 (2016) 12–33

33

Biggs, J. B., & Collis, K. F. (1982). Evaluating the quality of learning: the SOLO taxonomy (structure of the observed learning outcome). New York: Academic Press. Bryan, L. (2003). Nestedness of beliefs: examining a prospective elementary teacher’s belief system about science teaching and learning. Journal of Research in Science Teaching, 40, 835–868. Bybee, R. (2013). The case for STEM education: Challenges and opportunities. Arlington, Virginia: NSTA Press, National Science teachers’ Association116. Crawford (2012). Moving the essence of inquiry into the classroom: engaging teachers and students in authentic science. In K. C. D. Tan, & M. Kim (Eds.), Issues and challenges in science education research: moving forwardNew York: Springer Science and Business Media. Cresswell, J. (2012). Educational research: planning, conducting, and evaluating quantitative and qualitative research, 4th ed. Boston: Pearson. diSessa, A. A. (2004). Metarepresentation: native competence and targets for instruction. Cognition and Instruction, 22, 293–331. Gillies, R. M., Nichols, K., & Khan, A. (2015). The effects of scientific representations on primary students’ development of scientific discourse and conceptual understandings during cooperative contemporary inquiry-science. Cambridge Journal of Education. http://dx.doi.org/10.1080/0305764X.2014.988681. Goldstein, H. (2005). Multilevel statistical models. London: Arnold, Holder Headline Group. Grigg, J., Kelly, K. A., Gamoran, A., & Borman, G. D. (2012). Effects of two scientific inquiry professional development interventions on teaching practice. Educational Evaluation and Policy Analysis, 35(1), 38–56. Hilton, A., & Nichols, K. (2011). Representational practices that impact on student conceptual understanding and representational competence in bonding chemistry. International Journal of Science Education, 33, 2215–2246. Howitt, C., Rennie, L., Heard, M., & Yuncken, L. (2009). The scientists in schools project. Teaching Science, 55(1), 35–38. Hubber, P., Tytler, R., & Haslam, F. (2010). Teaching and learning about force with a representational focus: pedagogy and teacher change. Research in Science Education, 40, 5–28. Kockelman, P. (2007). The relation between meaning, power, and knowledge. Current Anthropology, 48(3), 375–401. Kozma, R., & Russell, J. (2005). Students becoming chemists: developing representational competence. In J. K. Gilbert (Ed.), Netherlands: Springer. Jaipal, K. (2010). Meaning making through multiple modalities in the biology classroom: a multimodal semiotics discourse analysis. Science Education, 94(1), 48–72. Lee, O., Hart, J. E., Cuevas, P., & Enders, C. (2004). Professional development in inquiry-based science for elementary teachers of diverse student groups. Journal of Research in Science Teaching, 41(10), 1021–1043. Lemke, J.L. (2004). The literacies of science. Retrieved from: http://jaylemke. squarespace.com/storage/Literacies-of-science-2004.pdf Accessed 20.06.07. Lieberman, A. (1995). Practices that support teacher development: transforming conceptions of professional learning. Phi Delta Kappan, 76(8), 591–596. McKinnon, M., & Lamberts, R. (2014). Influencing science teaching self-efficacy beliefs of primary school teachers: a longitudinal case study. International Journal of Science Education, 4(2), 172–194. McLean, C. A., & Rowsell, J. (2013). (Re) designing literacy teacher education: a call for change. Teaching Education, 24(1), 1–26. Nichols, K., Hanan, J., & Ranasinghe, M. (2013a). Transforming the social practices of learning with representations: a study on disciplinary discourse. Research in Science Education, 43, 179–208. Nichols, K., Ranasinghe, M., & Hanan, J. (2013b). Translating between representations in a social context: a study of undergraduate science students’ representational fluency. Instructional Science, 41, 699–728. Nichols, K., Gillies, R., & Hedberg, J. (2015a). Argumentation-based collaborative inquiry in science through representational work: impact on primary students’ representational fluency. Research in Science Education. http://dx.doi.org/10.1007/s11165-014-9456-4. Nichols, K., Stevenson, M., Hedberg, J., & Gillies, R. (2015b). Primary teachers’ representational practices: from competency to fluency. Cambridge Journal of Education. http://dx.doi.org/10.1080/0305764X.2015.1068741. Osborne, J. F., Simon, S., & Collins, S. (2003). Attitudes towards science: a review of the literature and its implications. International Journal of Science Education, 25, 1049–1079. Prain, V., & Waldrip, B. (2006). An exploratory study of teachers’ and students’ use of multi-modal representations of concepts in primary science. International Journal of Science Education, 28(15), 1843–1866. Posnanski, T. J. (2002). Professional development programs for elementary science teachers: an analysis of teacher self-efficacy beliefs and a professional development model. Journal of Science Teacher Education, 13(2), 189–220. Rabe-Hesketh, S., & Skrondal, A. (2005). Multilevel and longitudinal modeling using stata. College Station, TX: Stata Press. Riggs, I. M., & Enochs, L. (1990). Toward the development of an elementary teacher’s science teaching efficacy belief instrument. Science Education, 74(6), 625–637. Rundgren, C.-J., Hirsch, R., Chang Rundgren, S.-N., & Tibell, L. A. E. (2012). Students’ communicative resources in relation to their conceptual understanding— the role of non-conventionalized expressions in making sense of visualizations of protein function. Research in Science Education, 42, 891–913. Schunk, D. H. (1991). Self-efficacy and academic motivation. Educational Psychologist, 26(3/4), 207–231. Supovitz, J. A., & Turner, H. M. (2000). The effects of professional development on science teaching practices and classroom culture. Journal of Research in Science Teaching, 37(9), 963–980. Tschannen-Moran, M., & Hoy, A. W. (2001). Teacher efficacy: capturing an elusive construct. Teaching and Teacher Education, 17, 783–805. Tseng, C.-H., Tuan, H.-L., & Chin, C.-C. (2012). How to help teachers develop inquiry teaching: perspectives from experienced science teachers. Research in Science Education, 43(2), 809–825.