Introductory astronomy and cosmology courses at large universities often enroll hundreds of students, most of whom are not science majors. Instructors know that writing deepens understanding, yet the grading burden usually pushes them toward multiple-choice exams. A popular science text such as J. Richard Gott III's "Time Travel in Einstein's Universe" can anchor a manageable writing component if the assignment is designed carefully.

The key is to limit scope. Instead of a long term paper, assign a series of short essays of around 500 words, each focused on one idea from the reading, such as how gravity can bend time or what the author means by a closed loop in spacetime. Short assignments are easier to grade, easier to give feedback on, and far less likely to be plagiarized from a single source.

Large courses also benefit from standardized prompts and a single shared rubric used by every grader. When teaching assistants interpret criteria differently, students in different sections receive inconsistent grades for similar work. A common rubric, supported by a short calibration meeting, goes a long way toward preventing that.

Train teaching assistants with real student samples

Graduate teaching assistants are usually strong scientists but may have little experience grading writing. Before the first assignment, have them score the same three sample essays independently, then discuss where and why their scores differed. This exercise reveals hidden disagreements about what counts as a thorough explanation, and it can be completed in under an hour.

• Provide one high, one middle, and one low sample essay for each assignment.
• Have every grader score the samples before seeing the answer key.
• Discuss any score gaps larger than one rubric level.
• Record decisions about borderline cases in a shared document.
• Repeat a mini calibration halfway through the grading period.

Focus feedback on the few ideas that matter most

With hundreds of papers, nobody can write detailed comments on every sentence. Choose two or three recurring issues for each assignment, such as accuracy of a key concept or clarity of explanation, and give feedback only on those. Students tend to remember focused feedback better than long lists of corrections.

A bank of reusable comments tied to rubric criteria can speed this process considerably. For instance, a prepared note might explain the difference between describing a phenomenon and explaining why it occurs. Graders can paste and personalize these notes, preserving quality while saving time.

Protect academic integrity without turning grading into policing

Popular books about physics are heavily summarized online, and students in large courses may be tempted to borrow phrasing. Designing prompts that require a personal application, such as relating a concept to something the student has observed, makes borrowed text easier to spot and less useful. It also produces more interesting essays to read.

Clear syllabus language about permitted tools and sources reduces confusion and conflict later. Students should know whether summaries, study guides, and AI tools are allowed, and how to cite them if they are. Setting those expectations early lets you spend grading time on learning rather than detective work.

Use technology to keep the workload sustainable

Even with short assignments, several hundred essays a week can exhaust a teaching team. Automating the first pass of rubric scoring and draft feedback can return days of time to instructors, particularly for criteria like organization and completeness. Human graders can then concentrate on the scientific accuracy that requires expertise.

The professors who succeed with this approach tend to treat AI as a support for their rubric, not a substitute for their judgment. They spot check results, adjust criteria after each assignment, and keep students informed about how feedback is produced. That transparency builds trust and keeps the focus on learning.