Free Resource

AI Prompt Cheat Sheet
for Researchers

19 ready-to-use prompts organized by research task. Copy, customize, and use immediately with Claude, ChatGPT, or any LLM.

📚Literature Review📊Data Analysis🏆Grant Writing✍️Paper Writing & Editing💻Code Generation
How to use:

Each prompt uses [PLACEHOLDERS] in brackets. Replace them with your specific details before pasting into your AI tool. The more specific you are, the better the output.

📚
Category

Literature Review

Initial Topic Scan

Literature Review #1
copy
I'm researching [TOPIC] in [FIELD]. Give me a structured overview of the key themes, major debates, and seminal papers I should know about. Format it as: (1) Core concepts, (2) Key open questions, (3) Landmark papers (with brief descriptions), (4) Recent developments.

What it does

Rapidly maps a new topic before diving into papers — great for entering an unfamiliar subfield or writing an introduction.

💡

Customization tip

Replace [TOPIC] and [FIELD] with specifics (e.g., 'mechanistic interpretability' in 'AI safety'). Add your current knowledge level so it can calibrate depth.

Paper Digest

Literature Review #2
copy
Summarize this paper for me: [PASTE ABSTRACT OR FULL TEXT]. Focus on: (1) Research question and motivation, (2) Key methodology, (3) Main findings and their significance, (4) Limitations the authors acknowledge, (5) How it relates to [MY RESEARCH AREA].

What it does

Extracts the essentials from a paper in under a minute — especially useful for screening dozens of papers quickly.

💡

Customization tip

Paste the abstract for quick triage, or the full text for deeper analysis. Specify your research area in the last point for relevance filtering.

Research Gap Finder

Literature Review #3
copy
Based on these papers I've reviewed: [LIST KEY PAPERS/FINDINGS], what appear to be the most significant gaps, contradictions, or underexplored areas in the literature? Rank them by likely impact and feasibility for a single research group to address.

What it does

Synthesizes multiple sources to surface original contribution opportunities — ideal when writing the 'gap in the literature' section.

💡

Customization tip

List papers with one-line summaries of their findings for best results. Ask it to focus on methodological, theoretical, or empirical gaps specifically.

Conflicting Findings Reconciler

Literature Review #4
copy
These two sets of studies reach opposing conclusions about [TOPIC]: Group A finds [X], Group B finds [Y]. Analyze the methodological, sample, and contextual differences that might explain this discrepancy. What would a synthesis position look like?

What it does

Helps you write nuanced literature reviews that go beyond citing papers — shows understanding of the field's tensions.

💡

Customization tip

Include specific paper details (sample sizes, methods, populations) for a more substantive analysis rather than generic hedging.

📊
Category

Data Analysis

Dataset First Look

Data Analysis #1
copy
I have a dataset with the following structure: [DESCRIBE COLUMNS, DATA TYPES, N ROWS]. My research question is: [QUESTION]. What exploratory analyses should I run first? What potential issues should I check for (missing data, outliers, distributional assumptions)? Give me a step-by-step EDA plan.

What it does

Creates a systematic exploratory plan before touching any analysis code — prevents costly mistakes downstream.

💡

Customization tip

Paste actual column names and a few example rows. The more concrete you are, the more specific (and useful) the plan.

Statistical Test Selector

Data Analysis #2
copy
I want to test whether [HYPOTHESIS] using data where my outcome variable is [TYPE: continuous/binary/count/etc.], my main predictor is [TYPE], and I have [N] observations. I [do/don't] have repeated measures. What statistical tests are appropriate? Walk me through the assumptions for each and how to check them.

What it does

Guides test selection based on data characteristics — reduces the chance of using an inappropriate analysis.

💡

Customization tip

Include any domain-specific constraints (e.g., clustered data, longitudinal structure, small sample size). Ask it to compare frequentist and Bayesian options if relevant.

Results Interpreter

Data Analysis #3
copy
Here are my statistical results: [PASTE OUTPUT — coefficients, p-values, confidence intervals, etc.]. My research question was: [QUESTION]. Help me interpret these findings accurately, flag any results I might be misreading, and suggest how to report them in APA/AMA/Nature format.

What it does

Translates raw statistical output into plain-language interpretations and catches common misinterpretations.

💡

Customization tip

Include the full model output, not just significant results. Ask for help distinguishing statistical significance from practical/clinical significance.

Analysis Code Reviewer

Data Analysis #4
copy
Review this [Python/R] analysis code for my study on [TOPIC]: [PASTE CODE]. Check for: (1) Statistical errors or inappropriate methods, (2) Potential bugs, (3) Missing data handling issues, (4) Reproducibility problems, (5) Ways to make it cleaner and more efficient.

What it does

Independent code review that catches statistical and programming errors before submission.

💡

Customization tip

Include comments describing what each section is supposed to do. Ask it to flag anything that would be flagged in peer review.

🏆
Category

Grant Writing

Specific Aims Drafter

Grant Writing #1
copy
Help me draft Specific Aims for an NIH [R01/R21/etc.] grant. My research focuses on [TOPIC]. The central hypothesis is [HYPOTHESIS]. My preliminary data shows [KEY FINDINGS]. The three aims I'm considering are: [AIM 1], [AIM 2], [AIM 3]. Write a one-page Specific Aims section with: opening hook, knowledge gap, central hypothesis, three aims with expected outcomes, and a closing paragraph on impact.

What it does

Generates a structured first draft of the most critical grant section — the Aims page that reviewers read first.

💡

Customization tip

Give it your preliminary data points as bullet points. Iterate on the opening hook — ask for 3 different options and choose the strongest.

Significance & Innovation Strengthener

Grant Writing #2
copy
Here is my draft Significance and Innovation section: [PASTE TEXT]. Strengthen it by: (1) Making the knowledge gap more concrete with specific citations where I've noted [CITE], (2) Sharpening the 'why now' argument, (3) Making the innovation claims more specific and defensible, (4) Ensuring alignment with [FUNDING AGENCY] priorities. Keep the same approximate length.

What it does

Sharpens the two sections reviewers use to score scientific merit — typically the biggest differentiator in competitive grants.

💡

Customization tip

Tell it the study section or program officer priorities. Ask it to flag any claims that a skeptical reviewer might challenge.

Reviewer Objection Anticipator

Grant Writing #3
copy
I'm submitting a grant on [TOPIC] with this approach: [SUMMARIZE DESIGN]. Act as a tough but fair NIH study section reviewer. What are the most likely critiques about (1) significance, (2) innovation, (3) approach feasibility, (4) investigator qualifications, (5) environment? For each critique, suggest how I should address it proactively in the application.

What it does

Stress-tests the application before submission — identifies weaknesses you can address rather than discover in reviews.

💡

Customization tip

Share your actual weaknesses (small sample, novel method, early-stage PI) so it can give targeted rather than generic advice.

✍️
Category

Paper Writing & Editing

Abstract Generator

Paper Writing & Editing #1
copy
Write a structured abstract for my paper on [TOPIC]. Background: [2-3 sentences on the problem]. Methods: [KEY METHODS]. Results: [KEY FINDINGS WITH NUMBERS]. Conclusions: [MAIN TAKEAWAY]. The target journal is [JOURNAL]. Follow their abstract format: [structured/unstructured, word limit]. Make the opening sentence a compelling hook for a researcher who doesn't yet know they care about this topic.

What it does

Drafts a polished abstract that matches journal requirements — saves hours of wrestling with word limits.

💡

Customization tip

Give it real numbers from your results. Ask for 2-3 versions of the opening sentence and pick the strongest. Check the journal guidelines and specify them explicitly.

Discussion Section Builder

Paper Writing & Editing #2
copy
Help me write the Discussion section for my paper. My main findings are: [LIST FINDINGS]. The expected findings from prior literature were: [EXPECTATIONS]. Surprising or novel findings: [SURPRISES]. Key limitations: [LIMITATIONS]. Discuss: (1) How findings support/challenge existing theory, (2) Mechanistic explanations for surprising results, (3) Practical implications, (4) Limitations with suggested future work. Match the tone of [TARGET JOURNAL].

What it does

Structures the Discussion logically and ensures all findings are interpreted, not just restated.

💡

Customization tip

Don't ask it to spin limitations away — a strong Discussion acknowledges them and frames them as future work. Add 'be honest about uncertainty' to the prompt.

Clarity & Conciseness Editor

Paper Writing & Editing #3
copy
Edit this paragraph from my [methods/results/discussion] section for clarity and conciseness without changing the scientific meaning: [PASTE TEXT]. Specific goals: eliminate passive voice where possible, cut words by ~20%, replace jargon with plain language where appropriate (audience: [expert/general scientific]), and fix any awkward constructions. Show me the revised version and a brief note on the main changes.

What it does

Tightens dense academic prose — particularly useful for methods sections and for non-native English speakers.

💡

Customization tip

Do this paragraph by paragraph rather than all at once — it's easier to review and you maintain more control over your voice.

Reviewer Response Drafter

Paper Writing & Editing #4
copy
Help me respond to this peer review comment: [PASTE COMMENT]. My response should: (1) Acknowledge what the reviewer correctly identified, (2) Explain what changes I made (or why I didn't change it), (3) Be respectful but firm if I disagree. Here's what I actually did in revision: [YOUR CHANGES]. Draft a response that is professional, concise, and shows the reviewer their concern was taken seriously.

What it does

Converts reviewer comments into structured, professional responses — reduces the emotional labor of revision.

💡

Customization tip

Use this for every comment, even ones you're rejecting. A well-reasoned 'we disagree because...' response is far better than silence.

💻
Category

Code Generation

Data Processing Pipeline

Code Generation #1
copy
Write [Python/R] code to process my research data. Input: [DESCRIBE FILE FORMAT — CSV, JSON, etc. with key columns]. I need to: (1) Load and validate the data, (2) Handle missing values by [strategy], (3) [SPECIFIC TRANSFORMATIONS], (4) Output a clean dataset for analysis. Include docstrings, error handling for malformed inputs, and a brief usage example.

What it does

Generates production-quality data preprocessing code with proper error handling — not just a quick script.

💡

Customization tip

Paste a few rows of sample data or describe the exact column names. Ask it to add assertions that verify data quality at each step.

Statistical Analysis Script

Code Generation #2
copy
Write [Python with statsmodels/scipy/pingouin OR R] code to run [ANALYSIS TYPE] on my dataset. Variables: [DESCRIBE]. I need to: (1) Check assumptions ([LIST ASSUMPTIONS]), (2) Run the main analysis, (3) Compute effect sizes, (4) Generate a results table suitable for a paper. Add comments explaining each step and flag where I should review the output manually.

What it does

Creates fully-commented analysis code that runs the test, checks assumptions, and formats output for papers.

💡

Customization tip

Specify the exact package you want (e.g., 'use lme4 in R' or 'use pingouin for Python'). Ask it to include a separate section for assumption checking.

Figure Generator

Code Generation #3
copy
Write Python (matplotlib/seaborn/plotly) code to create a publication-quality [FIGURE TYPE] for my paper. Data: [DESCRIBE STRUCTURE]. The figure should: (1) Show [MAIN COMPARISON/PATTERN], (2) Use colorblind-friendly colors, (3) Include error bars showing [SEM/SD/CI], (4) Be exported at 300 DPI as both PNG and SVG. Target journal: [JOURNAL] — apply their style guidelines if you know them.

What it does

Generates figures that meet journal submission standards — colorblind-safe, high resolution, properly labeled.

💡

Customization tip

Mention the journal's preferred color palette if you know it. Ask it to also generate a version with a white background for preprints vs. a journal-specific version.

Simulation / Synthetic Data Generator

Code Generation #4
copy
Write Python code to simulate [DESCRIBE EXPERIMENT/DATASET]. I need synthetic data with these properties: [DISTRIBUTIONS, CORRELATIONS, SAMPLE SIZES, EFFECT SIZES]. Use this for: (1) Power analysis, (2) Testing my analysis pipeline before real data is collected, (3) Checking if my statistical approach can recover the true effect. Include a seed for reproducibility and a function to vary key parameters.

What it does

Creates parameterized simulations for power analysis and pipeline testing — standard practice before collecting expensive data.

💡

Customization tip

Base effect sizes on published literature in your area. Ask it to also generate a 'null condition' (effect size = 0) for testing Type I error rates.

Go Deeper

Want personalized prompts and workflows for your specific research?

These prompts are a starting point. In a 1:1 consultation, we build custom prompt libraries and AI workflows tailored to your exact methodology, domain, and tools — saving you hours every week.

Book a Consultation

Or email hello@labwise.ai with questions.