Interpreting Anomalous Primary Data in the Laboratory: Findings From a Longitudinal Study

• In Event: Promoting Epistemic Practices in Science Classes

Abstract

Objectives and theoretical framework
Our study is framed by literature on epistemic cognition (Chinn et al., 2011; Kelly 2008; Pluta et al., 2011) and by policy recommendations about students' participation in scientific practices (NRC, 2012). The practices examined are analyzing and interpreting data and drawing conclusions. Compared to studies about modeling and argumentation, students' engagement in scientific practices when carrying out investigations has received less attention. The longitudinal study has the goal of examining progress in students' enactment of epistemic practices in the laboratory. This paper focuses on students' responses to anomalous data, drawing from Chinn & Brewer (1993; 1998; 2001). Few studies (Lin, 2007; Shepardson & Moje, 1999; Toplis, 2007) address responses to anomalous data in experiments. Our objective is to identify patterns in students' responses to primary data, which in contrast with use of provided secondary data is an understudied issue (Kanari & Millar, 2004).
Methods and data sources
A qualitative longitudinal study was conducted during two years, 9th and 10th grade. (21 students in the 9th grade; 10 students in the 10th grade, when science is optional). A set of five problem-based learning (PBL) laboratory tasks was designed. Students were required to plan and carry out experiments in order to solve problems set in authentic contexts. In coherence with the PBL approach scaffolding were designed to support students, for instance in making their reasoning explicit. Data collected include video and audio recording, written productions and interviews. A rubric was elaborated drawing on Chinn & Brewer's (2001), adapting it for experimental contexts.
Selected results
Unlike studies about theory change, the occurrence of anomalous data was not embedded in the design, but resulted from low manipulative skills. It occurred in over a half of cases across small groups and tasks. 1) Our findings suggest that primary data evaluation in a school laboratory context relates to confirmation or rejection of predictions about experimental results, rather than to comparing alternative theories; 2) A combination of different responses to data conflicting with predictions was found, which agrees with Lin (2007); 3) The most frequent response was reinterpretation of data, with particular students taking this role; 4) We suggest that in laboratories, reinterpretation should be split in: a) interpreting data as methodological errors (which was true, so this was not coded as rejection) and suggesting potential reasons; b) interpreting data as valid and suggesting alternative explanations; 5) Two categories of responses found in our study could be added: a) proposals of repeating the experiment in combination with 4a; b) Implications drawn for real life, specific for authentic contexts, with 4b; 6) While in conceptual change contexts, theory change is the desired achievement; in laboratory contexts where anomalous data may result from errors, reinterpretation of data, and elaboration of explanations about their causes, may be a more sophisticated response.
Significance
The range of responses of individual students, the processes of justifying them in small groups, and their progress across the two years may contribute to insights into students' reasoning about anomalous data.

Authors

• Maria Pilar Jimenez-Aleixandre, Universidad de Santiago de Compostela

• Beatriz Crujeiras, University of Santiago de Compostela