Scientific explanations are important for students to construct their understanding scientific knowledge and link the knowledge with evidence. It is vital to instructionally support and measure the growth of students’ explanations in teachers’ daily practice. The purpose of this study is to explore methods to model the general growth trajectory of students’ learning of scientific explanations as well as its variability. Using the student notebook scores on four investigations within a physical science unit, a series of three-level hierarchical linear models will be conducted to identify the learning trajectory of student learning in the context of inquiry-based instructions. We will examine the mean and variability of the individual and group growth curves on students’ proficiency in scientific inquiry. Finally, we will discuss the issues related to methodological challenges and educational implications regarding to inquiry-based instructional and assessment practice.  

Research Aim

A recent NRC (2006) report suggests to focus on conducting research about children’s learning progressions in science. Learning progression is a way to describe the successively sophisticated ways of thinking as children learn about and investigate a topic over a span of time. It depends on instructional practices if it is to occur. In this paper we focus on learning about the development of students’ competence on one fundamental activity in scientific inquiry, the construction of scientific explanations. More specifically, we analyze the quality of students’ explanations over the course of the implementation of twelve investigations focusing on the buoyancy. Using the scores of students’ explanations and the level of understanding reflected on four critical investigations, we perform a series of analysis to identify the individual and group learning trajectories of their learning reflected by scientific explanations.

Coding Scientific Explanations

We used students’ science notebooks as the main source of information to analyze the characteristics of students’ explanations. The rationale is that notebooks are generated during the process of instruction and somehow reflect what students do in their classrooms (Warren-Little, Gearhart, Curry, & Kafka, 2003).

Coding focused on three aspects: (1) Quality of Explanations according to its three components (Claim: statement that answers a scientific problem; Evidence: data that supports a claim, and Reasoning: justification that shows why the data counts as evidence to support the claim; Kenyon & Reiser, 2006; Tzou, 2006). (2) Quality of Communication focusing on characteristics that have been considered important in students’ scientific communications, and (3) Student’s Level of Understanding focusing on students’ understanding about density based on a conceptual progress trajectory of density. For each aspect, we developed specific criteria to capture students’ level of explanations, communication, and understanding (see Table 1 for examples of coding claim and evidence).

Table 1. Examples of Questions to Score Quality of Scientific Explanations.

Method

Participants and Curriculum. Twelve middle school physical science teachers and their students participated in the study. All teachers implemented the Foundational Approaches in Science Teaching (FAST) middle-school science curriculum, an interdisciplinary, scientific inquiry program (Pottenger & Young, 1992). For this study, we chose to focus on four investigations from the first unit of FAST, Properties of Matter, which supports students in the development of a relative-density based explanation of sinking and floating.

Sources of Information. Teachers were asked to provide their students’ science notebooks at the end of the school year. Within each classroom, nine students’ notebooks were randomly selected from strata based on the students’ scores on the multiple-choice test administer at the end of the unit (three top-, three medium-, and three low-proficient).

Using an Access computer program, each student notebook entry was analyzed on six aspects based on the recommendations from the FAST curriculum materials: Problem, Vocabulary, Background, Method, Reporting Results, and Conclusions. Twelve of the 108 students’ notebooks were scored independently by three well-trained raters with an agreement at 78% for student performance. In this study, students’ notebook scores of the Conclusions aspect were used to model student learning of scientific explanations over the four investigations.

To validate the modeling, we also administered an open ended-question at the end of each of the four investigations to evaluate how well students understood the main ideas of those investigations (i.e., why do things float or sink?)

Data Analysis. Various methods, including Hierarchical linear modeling (HLM), will be used to analyze the longitudinal data over the four investigations in order to identify the latent growth curve of student understanding. The advantage of fitting the repeated measures with HLM is the flexibility to deal with the unbalanced data structure, such as repeated measures data with fixed measurement occasions where the data for some individuals is incomplete. This is the case in our data set since some teachers did not include some investigations in their instruction or some students did not complete their notebook entries due to absence.

Findings

Data analyses will focus on exploring ways to identify students’ learning trajectory of scientific explanations. First, we will present descriptive analysis of notebook scores around multiple aspects of students’ understanding. Second, we will run a set of analyses to identify a multi-level linear growth curve model for each aspect with longitudinal notebook scores nested within individuals nested within 12 classes. Third, we will validate the growth curve models using student scores from the repeated open-ended question. Finally, we will examine the patterns of those growth curves in relations to the instructional prompts and assessment supports (e.g., feedback or scoring rubrics that teachers used) to understand the instructional supports for promoting students’ understanding.

Based on the results, conclusion and discussion will focus on the general learning progress in students’ construction of scientific explanations as well as the variability on individual students’ learning of scientific explanations. We will also provide suggestions on the effects of teachers’ instructional and assessment supports variables and the implications of our findings in teachers’ professional development.

Theroretical and Educational Significance of the Research

This study will illustrate the methods of modeling the learning trajectory of students’ construction of scientific explanations in science notebooks. Furthermore, it will provide information on what instructional and teacher feedback practices observed seemed to be the most effective in supporting students’ science inquiry. The findings in this study will inform the science education community about modeling and measuring students’ learning progress in constructing scientific explanations. This study will also lead to a set of suggestions of relevant areas for teachers’ professional development.