1. Which prompts are best for standardised test preparation?
Sections 1 (argumentative) and 7 (compare and contrast) most directly develop the skills tested in standardised academic writing assessments. The argumentative prompts develop the claim-evidence-reasoning structure that most standardised tests reward. The comparative prompts develop the analytical framework that compare-contrast essay questions require. Regular practice with timed responses to these prompts is the most effective preparation for any standardised writing assessment.
2. Can these prompts be used across different subject areas?
Yes, and that cross-disciplinary use is one of the most valuable ways to use them. The analytical prompts in section 2 can be applied to science, history, literature, or economics depending on what content the student substitutes. The research prompts in section 5 work across all disciplines that use evidence. The problem-solving prompts in section 9 are particularly effective in science, engineering, and social studies contexts.
3. How long should responses to these prompts be?
For middle school students, 250 to 400 words is appropriate for most prompts. For high school, 400 to 700 words. For university level, 600 to 1000 words or more for the research and analytical prompts. The most important length guideline is that the response should be as long as the thinking requires — no padding to hit a word count, no cutting before the argument is complete. Quality of reasoning matters more than meeting a length target.
4. How can Claude help students with these prompts?
Claude is most useful for students at two stages: before writing, to explore the topic, identify counter-arguments, and stress-test a proposed thesis; and after a first draft, to get specific feedback on where the argument is weakest, where evidence is missing, and where the writing would benefit from greater precision. Claude does not write for students — it thinks with them. The distinction matters because the thinking is the skill being developed, and outsourcing the thinking to AI produces no learning.