This project builds off previous research in automatic text scoring and three dimensional (3D) learning progressions for high school science.

Research from automatic text scoring; Briefly, our developed tools apply techniques from artificial intelligence (AI), like natural language processing (NLP) and supervised machine learning (ML), to train a computer model to ‘score’ student constructed responses just like an expert would. Then these models can be used to classify additional student responses. Of course computers do not understand student writing like a human, but through AI computers can learn to identify and categorize writing like an expert would. Our tools use NLP to extract terms and phrases from student responses and use these as variables in ML classification algorithms. This forms the basis of the Constructed Response Classifier tool developed by the AACR research group in previous research studies.  We plan to leverage this existing tool in this project by training it to recognize different 3D assessment tasks and student responses.

As part of prior work, we developed and validated two NGSS-aligned LPs focusing on student ability to integrate the relevant NGSS dimensions (including DCIs, SEPs, and CCCs) when explaining phenomena involving electrical interactions. The LPs describe students’ increasing competency to develop causal models and explanations of phenomena involving electrical interactions focusing specifically on applying Coulomb’s law and Energy to explain electrical interactions at the macroscopic and atomic molecular level. The two LPs are closely related and follow similar logic. They both start from level 0 reflecting no substantive evidence of using either Coulomb’s law (for LP on Coulomb’s law, or Energy (for LP of Energy) to explain electrostatic phenomena. Further, level 1 reflects the ability to use Coulomb’s law and Energy for developing partial causal models and explanations of electrostatic phenomena. Level 2 reflects the ability to use either Coulomb’s law or Energy to explain electrostatic phenomena at the macroscopic level and partially at the atomic-molecular level. Finally, level 3 reflects the ability to use both Coulomb’s law and Energy ideas to explain electrostatic phenomena at the atomic molecular level. Table 1 shows a brief description of the levels for both LPs and reflects the features important for the current study. Both LPs focus on SEPs of Developing and Using Models and Constructing Explanations; the CCC of Cause and Effect and DCIs of Relationship between Energy and Forces and Types of Interactions.

 

3DLP1

3DLP 2

 

References: 

 

Conference Posters and Talks