Exploring Shifts in Middle School Learners’ Modeling Activity While Generating Drawings, Animations, and Computational Simulations of Molecular Diffusion

Abstract

Modeling and using technology are two practices of particular interest to K-12 science educators. These practices are inextricably linked among professionals, who engage in modeling activity with and across a variety of representational technologies. In this paper, we explore the practices of five sixth-grade girls as they generated models of smell diffusion using drawing, stop-motion animation, and computational simulation during a multi-day workshop. We analyze video, student discourse, and artifacts to address the questions: In what ways did learners’ modeling practices, reasoning about mechanism, and ideas about smell shift as they worked across this variety of representational technologies? And, what supports enabled them to persist and progress in the modeling activity? We found that the girls engaged in two distinct modeling cycles that reflected persistence and deepening engagement in the task. In the first, messing about, they focused on describing and representing many ideas related to the spread of smell at once. In the second, digging in, they focused on testing and revising specific mechanisms that underlie smell diffusion. Upon deeper analysis, we found these cycles were linked to the girls’ invention of “oogtom,” a representational object that encapsulated many ideas from the first cycle and allowed the girls to restart modeling with the mechanistic focus required to construct simulations. We analyze the role of activity design, facilitation, and technological infrastructure in this pattern of engagement over the course of the workshop and discuss implications for future research, curriculum design, and classroom practice.

Appendices

Appendix 1

See Table 5.

Table 5 The table presents the coding scheme used in study, including the categories, descriptions of the categories, and citations of literature that supported the a priori establishment of reasoning and practices codes

Appendix 2

Examples of coding disagreement (Hammer and Berland 2013).

Here, we describe three main types of systematic disagreement that emerged during analysis.

Type 1) Describing Phenomenon vs Setup Conditions. One frequent disagreement was between identifying reasoning about mechanism as Describing Phenomenon or Defining Setup Conditions. For example, the quote “We were deciding whether the skin of the orange smelled more pungent than the actual fruit part.” was coded by one author as Describing Phenomenon, since Eileen was recalling a general exploration, but coded by another author as Setup Conditions, since Eileen highlights the skin and flesh of the orange as different potential setups of the model. We preserved these disagreements because it might be unclear even to learners whether a particular noticing about the phenomenon will yield explicit selection of model components.

Type 2) Entailments of Setup Conditions. If a coder identified an exchange as involving Setup Conditions rather than Describing Phenomenon, they were also more likely to subsequently identify Entities & Properties, Behaviors or Interactions for the same code. For example, if a coder identified Eileen’s quote above about skin and fruit as Setup Conditions, they may subsequently code references to skin and fruit as Entities & Properties (the orange as peeled or unpeeled) of the model. We preserved these disagreements as evidence of the messiness of elaborating, articulating, and problematizing aspects of the phenomenon to be modeled.

Type 3) Representations as Evidence. There was some disagreement over whether participants did or did not reason about Behaviors or Interactions during a given video segment. Often, these disagreements had to do with whether the coder considered evidence from participants’ representational artifacts. For example, one group of girls placed a series of pipe cleaners emitting from an orange and pointing toward a nose in their animation. The group never verbally articulated why they did this, but the animation showed smell particles traveling in the direction they were pointing. One author used the animation as evidence for the codes Representation of Entities and Behavior—of the smell particles and their movement. Another who relied on the transcript only coded for Representation of Entities, but not their behavior. We preserved these disagreements because coders did not always have access to what participants did physically, and because coding participants’ representations without evidence from participant talk is necessarily interpretive.