Initially, we developed some items that were designed to simultaneously elicit multiple sections of the EBR Framework, thereby providing a more holistic picture of how students go through the entire reasoning process. An example of such an item that had been used in previous studies is, “Explain below why things sink and float. Write as much information as you need to explain your answer. Use evidence and examples to support your explanation.” Pilot testing showed that items like these failed to produce responses that could be interpreted consistently using the EBR Framework. Even with moderate scaffolding, such items elicited a wide variety of responses based on different parts of the framework; different responses might describe: (a) a rule by itself; (b) a rule supported by data with no evidence; (c) evidence and a premise but no rule, and many other combinations of components. Because different students responded using different parts of the framework, the quality of the responses could not be directly compared.
Pilot testing made it clear that an optimal items design required items that focused on a small section of the framework. Specifically, we determined that the item stem should present a single component and require the application of a single process to produce or evaluate an adjacent component. For example, several of the items we developed presented the student with data and asked them to summarize the data. In doing so, they would describe a piece of evidence via analysis (see Figure 1).
Because individual items using this template do not provide opportunities to observe students engaged in extended reasoning, we clustered items into sequences of three or four in which later items in the sequence build on the student’s responses to previous items. An example of an abbreviated item sequence is shown in Table 1. This sequence contains two items, Items 3a and 3b, in which the second item explicitly references and builds on the student’s answer to the first item. First, the student is presented with a claim (consisting of three observations) and asked in Item 3a to describe a premise that could be applied to explain the claim. To answer this