How Quill Gives Feedback

Here at Quill, we believe that offering feedback on student work is an essential part of the teaching process. Our feedback model, which has been recognized by several leading tech and education publications, is distinct among writing instruction tools because our software recognizes a wide range of errors and responds with instant, targeted feedback. Targeted feedback allows students to learn from their mistakes rather than simply memorize the right answer. Our software also shows students multiple correct answers when appropriate, which reinforces the idea that there may be several correct ways to write a sentence. Students can then decide which response they prefer and incorporate it into their writing.

In this article, we’re going to look underneath the hood at how Quill generates this immediate, targeted feedback for students.

Instant Feedback

Quill’s feedback encourages a robust revision process. Correct student responses cannot contain grammatical errors of any kind, regardless of the activity's grammar target. For example, if a student fails to add a punctuation mark at the end of a sentence but has done everything else correctly, they will be prompted with the feedback “Check for missing punctuation.” The student will move on to the next question once they rewrite the sentence with the correct punctuation included or after providing five incorrect answers.

After a student completes an activity, the teacher can go to the Activity Analysis report to view the specific grammar concepts that students have mastered or not yet mastered. This teaches students to incorporate a wide range of sentence construction strategies into their writing while allowing teachers to monitor student progress on key grammar skills.

Clear Guidance

Quill strives to empower students to become independent writers. Therefore, our feedback is designed to guide students as they work out the correct answer for themselves. To this end, we avoid using overly technical jargon for grammar concepts and instead explain each grammar concept with simple and clear language that is easily understood by students at many levels. Our feedback also guides students toward the correct answer by asking them to think about the structure and purpose of the sentence. In the example below, you can see what happens when a student forms a compound sentence using the joining word, and, instead of the more accurate joining word, so.

​​ Notice that we do not use the technical term, conjunction, but instead use the simpler term, joining word, to explain the error. By suggesting that the joining word should show a causal relationship, we show students how to make their sentence even more precise. This is especially helpful in cases when students structure their sentence in a way that sounds correct, but would be awkward or incorrect in mature writing.

Mastery-Based Grading

Another important feature of Quill’s feedback is its mastery-based grading system. With the Activity Analysis report, teachers can see the answers students submitted, the feedback they received, and the concepts they either have mastered or not yet mastered for each activity. Students can score in the Proficient (Green) proficiency level for scoring between 80-100%, in the Nearly proficient (Yellow) proficiency level for scoring between 60%-79%, or in the Not yet proficient (Red) proficiency level for scoring between 0%-59%. Our grading system gives students the chance to submit five multiple incorrect responses before they are prompted to move on, and for each incorrect submission, students are given more feedback that guides them toward the correct answer. Furthermore, students never see their numerical grade; rather, they can see their proficiency level based on how well they have mastered the concept.

Advanced Error Detection

All of this is possible because of Quill’s advanced error detection system. Our team works behind the scenes to ensure that each question is accurately graded with direct and relevant feedback. When grading questions, we begin by inputting all of the possible correct responses. This informs the algorithm with information about how to recognize when a student submission does not match the optimal response, and in this case, the algorithm will flag the submission as incorrect.

The algorithm can detect many kinds of errors, such as when a student includes a typo or forgets to capitalize the beginning of a sentence. Furthermore, the errors that students make are grouped together based on their respective grammar concepts and assigned a concept tag. Concept tags are used both to tell the computer the kind of feedback a student should receive and to track students’ mastery of specific grammar concepts across activities. The student will be provided with targeted feedback, and the teacher will see which grammatical concept the student missed.

For more information on Quill’s grading system, please see the following article.

Have more questions?

Please feel free to send a message to the Quill team using the green message bubble on the bottom right corner of the screen or email us at support@quill.org.