Modeling in NGSS

Jonathan Corbett
13 March 2016
Modeling has become a central activity in my classroom, and it has quickly become one of my favorite activities. I have students make models for pre-teaching formative assessments, mid-unit formative assessments, and end-of-unit summative assessments. Having students save their initial models -- say, models of the atom prior to any instruction -- for review and revision at the end of the unit, and the results are always informative and entertaining.
The first model of any science class is usually one of students' choosing. Popular topics in my classes have included a model of kicking a soccer ball, how an X-Box controller works, or the daily operations of a restaurant. The last on this list is my go-to example for students who feel "stuck". A birds-eye view of a restaurant tends to work best, and students can describe the relationships between the waitstaff, guests, kitchen, and owner. The flow of food and dining equipment and money work well for describing the relationships between components in the system.
Figure 1.
A sample model on chemical bonding completed by a ninth grade integrated physical science student. The model clearly demonstrates student understanding, but also shows a strong preference for written/essay responses over graphical representations.
Evidence is absent from this initial exercise. I usually provide them with the entire rubric, but have them cross the section off. The first modeling activity provides me feedback about student interest, helps me identify students who may excel (or have difficulty) with visual representations of systems, and who may need additional support in the planning process of making a model (see Scaffolding and Differentiation below).

Diagrams and Relationships​

Central to any model is some sort of diagram. Students seem most comfortable with this aspect, and often fail to distinguish between a drawing or diagram and a model. Explicit instruction, including examples, is needed before this difference is understood. Importantly, models include explicit descriptions of relationships between components and variables in a model. Arrows and symbols help show Structure/Function, Cause/Effect, and other NGSS CCCs.  

Additionally, many models include more than one diagram, including hypothetical situations, counterexamples, and details of the evidence used to develop the model. Finally, a proper model includes a figure caption that describes the system being shown.

Here is an example from the introductory lesson on modeling that I do with nearly all of my science classes. This is the teacher-created sample, and is very simple. Here is a rather excellent example made by a student in that same class.

Evidence

The NGSS practice of arguing from evidence can be combined in student models. I require students to provide the key lines of evidence that have been used to develop the model being produced.

This evidence often comes from Aperture Lab activities done in class. A sketch of a graph, a description of qualitative data or everyday observations, or a diagram of an experimental design often suffice in this category. For instance, the Alpha Particle Scattering Experiment would be used as evidence of the claim that "most of an atom is empty space". An explicit connection must be made between the evidence reviewed and the claim being made by the model.

Predictions

The quality of any model is its ability to make predictions. Both as a scientific exercise, and an opportunity to develop higher-order thinking in students, students are required to make explicit predictions using the models they generate. This may include predictions of experimental results, or counterexamples or applications to novel situations that students have not yet encountered. This component has led to some of my proudest moments as a teacher, and model predictions give real insight into the depth of a student's grasp of material. As an example, a freshman modeling the structure of an atom stumbled across the mechanisms for chemical bonding before ever being exposed to the concept in the curriculum.  Note that this model contains many inaccuracies. This assessment came at a time when only the most basic structure of the atom was being explored in class, approximately 1/4 of the way through the unit. Despite this, her insight into the structure of the atom is clear, and she would still Exceed Standard in the Predictions category of the rubric (below).

One important detail of predictions that students sometimes struggle with is the fact that the results must be explicit and clear. Students have often concluded something along the lines that "if the results were different, our model would be different, and we wouldn't know as much". Students should be taught that vague predictions such as these do not give evidence of understanding, and will not be seen as evidence of Meeting Standard.

Clarity

I often struggle with categories such as "neatness" in academic assignments. While neatness is important, a grade should reflect understanding of content, not clarity of handwriting.

Still, good modeling technique demands that a model be informative to the reader and clearly illustrate the system being shown. Further, like any assessment, the spirit is to provide evidence of understanding. For this reason, the clarity element was added to the modeling scoring guide.

I have found it useful to teach that models are meant to explicitly communicate what is known about a system, and that assumptions should always be minimized. Thus, students should not assume that I know the parts of an atom, or the relationship between the width of a water basin and the volume of water that it can hold. This attention to detail requires that students make their thinking concrete on paper. Initially, students respond with something along the lines of "you knew what I meant". The response to this is that I score assessments based on the evidence I was provided. Just like in science, the observer should have a minimal impact on the results; any other science teacher should be able to pick up my rubrics and my students' models and reach nearly the same conclusions about the students' understanding.

Evaluating Models and Models as Assessment

The majority of my assessments now include some element of modeling. These are not always complete models, but many are. The NGSS Evidence Statements provide a very clear scaffold around which modeling activities can be produced. Please see the page on Standards-based grading for more on how I use this rubric.
Generic Modeling Rubric
Students are always provided a copy of the modeling rubric when working on any model -- whether it is a pre-model or a final assessment. Students are also asked to self-asses their models prior to submitting them to the teacher.

Scaffolding and Differentiation

By their nature, modeling assessments are easily differentiated. Students can be provided a minimal scaffold (right), or almost any degree of partially completed diagram. For high differentiation, evidence statements can be provided in multiple choice or fill-in-the-blank format. Word banks of diagrams, relationships, and/or evidence can also be provided. 

The evidence section also directs students focus to the model's ability to make predictions. This is often by means of providing evidence that would refute the model, or predicting how the model would be different if an experiment had yielded different results. These scenarios can be provided to students who would benefit from upward differentiation.
Figure 2.
Initial version of the modeling prototype.