ThinkCERCA's Award-Winning, Research-Based Automated Feedback and Scoring

ThinkCERCA’s Award-Winning, Research-Based Automated Scaffolds

Automated reading scaffolds, feedback, and scoring are continuously optimized feature of ThinkCERCA's platform. Access to these features is customized based on district and school policies, with additional teacher-controlled features that allow teachers to turn off these scaffolds as needed.

Among hundreds of competitors in a global competition hosted by The Learning Agency, ThinkCERCA won the 2022 assessment competition with our writing feedback tool. In 2024, the Learning Agency awarded ThinkCERCA a second grant-funded opportunity to optimize our feedback and scoring tools.

Feedback vs. Autoscoring

AI Feedback, which acts like a writing coach to help students strengthen their arguments with guidance on specific ways to improve evidence-based writing (claims, evidence, reasoning, and counterarguments)

Autoscoring automatically evaluates student writing using our rubric and provides a score along with helpful, natural language explanations. AI Feedback is designed for students, while Autoscoring supports both students and teachers.

We utilize OpenAI's LLMs via API to power our AI Feedback and Autoscoring features, which support student writing development and assessment.

We also implemented Immersive Reader to support students with additional reading scaffolds.

• Pre-teaching lessons: To fill background knowledge gaps and rehearse skills
• Text Customization: Adjust color, size, and spacing
• Syllable Splitting: For fluency practice/deeper analysis
• Parts of Speech: For deeper analysis
• Picture Dictionary → Visual representation of words
• Line Focus → To reduce distraction
• Text-to-Speech → Adjustable voices for read-aloud
• Adjustable Reading Speed → To slow down or speed up the read-aloud
• Real-time Translation → Read in 170 languages and leverage bilingual read-aloud in more than 50 languages
• Speech-to-Text: To enable drafting and revision by voice in multiple languages

The goal of these tools is simple: save teachers' time and enable student learning.

Controlling Access

District Level Feedback and Scoring Settings

District partners align with ThinkCERCA's Success team to determine the best default settings for their community. The district can allow schools or teachers to determine whether or not to leverage feedback and scoring tools at the assignment level.

District and school-level leaders can also determine the policy around scoring. The feedback and scoring scaffolds provide real-time personalized feedback for students, and our students are motivated to engage in the revision process as a result of using the tool, with an average of 2.24 scans per writer. Therefore, most schools take full advantage of that feature, except in higher-stakes formal and informal assessment settings.

Automatic scoring should never automatically populate gradebooks; users can determine whether students can preview the scores, which are clearly marked as CERCA AI, before or after teacher review. Once a teacher has reviewed, edited, and approved the CERCA’s automated comments and score, the score can be exported to the LMS or manually recorded in gradebooks. Teachers can change the setting when assigning the lesson. Later, if needed, teachers can turn it on and off in the lesson detail page.

As always, settings can be adjusted by the district to match district- and school-level policies.

Teachers can always remove these features for formal and informal assessment at the time of individual lesson assignment.

Continuous Evaluation and Improvement Processes

ThinkCERCA's award-winning, research-based feedback is continuously improved based on correlation studies and feedback from professional writing instructors who evaluate tens of thousands of papers annually. We also review student and teacher feedback collected on the platform. Additionally, through classroom observations as well as user interviews and surveys, ThinkCERCA's team continuously reflects on the impact on student outcomes.

Ethical Use of Technology, Including Automated Feedback

Teachers cannot be replaced by technology. Peer discourse cannot be replaced by technology. In fact, humans are the main reason why schools and classrooms have endured for thousands of years and will persist despite technological innovation.

From day one, ThinkCERCA has promoted a blended learning model where teachers, peers, print materials, and technology are optimized to support, enable, and motivate student learning.

Providing time for teachers to conference and enabling students to access the help they need to catch up, keep up, or get ahead is the goal. After all, discourse-rich classrooms is where learning happens.

The added bonus is that these scaffolds support the development of transferable skills for teachers and students.

Students learning to manage their use of scaffolds is an important aspect of young adult success. Teachers learning how to provide nurturing, actionable feedback while holding the same high expectations for all students is one of the many benefits of our award-winning technology.

Data Security and Privacy Protocols

We take student privacy and data security very seriously in our AI-powered educational features. Here are the key details regarding our approach:

No personally identifiable information (PII) is ever transmitted to external AI services

Only essential academic content is shared - reading passages, writing prompts, student responses, and instructional content.

Student writing is never used to train external models - our agreements with OpenAI ensure that inputs and outputs from our platform are not used to train their models

All AI inputs and outputs are stored securely within our internal systems using industry-standard encryption and access controls.