Developing Effective Research Questions

How I Cut My Legal Research Time in Half (Without Lowering Quality) In law school, I used to spend hours researching cases, scrolling through long judgments, and struggling to find the right precedent. Then, I discovered something—technology can do half the work for you. Here’s how I started using tech to improve my legal research efficiency (and how you can too): ➡ I stopped relying only on Google and SCC At first, I used SCC and Google like everyone else. But then I explored AI-powered tools like CaseMine, Manupatra’s AI assist, and LexisNexis search filters. These tools don’t just show cases—they analyze patterns, suggest related cases, and even highlight the most relevant paragraphs. ➡ I used AI tools to summarize long judgments Instead of reading 100+ pages of a judgment, I used AI tools like Judgment Summarizer (Judi.AI), ChatGPT, and Casetext’s CARA to get quick summaries. I still cross-checked the key paragraphs, but this saved me hours of skimming through irrelevant sections. ➡ I automated citations instead of doing them manually I used to format citations manually (which was painfully slow). Then I found tools like Zotero, Refworks LLC, and EndNote, which automatically generate and format case citations in Bluebook, OSCOLA, or any other style. ➡ I learned how to use Boolean search effectively Most students waste time searching with plain keywords. I learned Boolean operators (like AND, OR, NOT, NEAR) to refine my searches. Instead of searching "arbitration clause invalid enforcement India", I used: 📌 “arbitration clause” AND (“invalid” OR “unenforceable”) AND India This pulled up precise, relevant results—faster and with less junk. ➡ I created a personal case law database Instead of searching for the same cases repeatedly, I started saving and tagging judgments using Notion, Microsoft OneNote, or Evernote. Whenever I found an important case, I stored it with key takeaways, so I never had to research it again. ➡ I used contract analysis software for drafting research For contract-related research, I used tools like Kira Systems and Lawgeex. These platforms analyze contracts and highlight risky clauses, giving me a head start before I even begin drafting. ➡ I practiced speed reading with tech tools Reading long judgments was slowing me down. So, I used speed-reading tools like Spritz Reader and Reedy to improve my reading efficiency, helping me absorb legal texts faster. ➡ I set up alerts for legal updates Instead of manually checking for new laws, I set up alerts on LexisNexis, SCC Online, and Google Alerts to notify me whenever new judgments or amendments were published in my areas of interest. The result? Faster research, more accurate results, and more time for actual analysis instead of just searching. If you’re still researching the old-school way, start using technology. Lawyers who use tech don’t just work faster—they work smarter.

What is a question worth asking? In academia, this is part of the basic researcher training and it never stops, not even as you get more senior. Indeed, most research projects do not fail because of bad methods. They fail because they start with the wrong question. Often, research questions are treated as something we quickly formulate before moving on to the "real" work. -> That is a mistake! A weak research question rarely becomes a strong study. It affects the theory, the methods, and the contribution. Conversely, a well-crafted research question can make many subsequent decisions much easier. That is why I was intrigued about half a year ago when my colleague Christian Tröster developed an AI-supported coaching tool specifically designed to help researchers sharpen their research questions. Needless to say, I tried it out for a few serious spins. What I particularly liked is that the tool does not simply generate answers. Instead, it challenges your assumptions, probes for tensions and puzzles, checks relevant literature, and helps you think more rigorously about what exactly makes a question worth asking. The tool is free to use and may be particularly helpful for students, doctoral researchers, and early-career scholars ... but I suspect many senior academics would also benefit from putting some of their ideas through the process. 👉 https://lnkd.in/exHn94bU (check it out) A nice aspect is that the coaching logic is grounded in decades of research on scientific inquiry, interestingness, theory development, and phenomenon-driven research rather than generic AI prompting. Sessions are stored anonymously (no personal data, EU servers, consent on the first screen, no one will “steal” your idea); the data feeds a research project behind the tool and its further improvement. Once it has a solid proof-of-concept, Christian will publish the underlying theoretical architecture and the code Fully OPEN ACCESS. If you give the tool a try, I'd be curious to hear what you think. Or let Christian know directly. --- As AI becomes increasingly embedded in scholarly work, I suspect tools like this will become part of the standard research workflow, not replacing scientific judgment, but helping us exercise it more deliberately. Getting better! For an open access article on the matter see here: https://lnkd.in/eUDFuPaa For other AI tools for the research process, see here: https://lnkd.in/eJQ4Y6uS

Three years ago, I was wrapping up my Master’s thesis: two years of research on convolutional networks’ interpretability. Back then, I was deep in both writing and research, crafting thesis drafts while also scripting YouTube videos to explain research papers. I lived in research-mode. And honestly? I wish tools like AnswerThis had existed. Formulating a thesis statement took me weeks. Then came revisions, more feedback, more rewriting... not to mention searching through hundreds of papers to find gaps and build a coherent narrative. All before the experiments section, built over 1.5 years of work. Still, writing was the hardest part. Connecting ideas clearly, crafting a cohesive story, and making it all "good." Today, tools can cut through that pain. Here’s how AnswerThis (link in the comments) can help you write a thesis statement 10x faster: 1. Start with your general topic, e.g., “AI Ethics.” 2. Head to the Structured Literature Review tool. 3. Adjust filters like citation count or date range to narrow down high-quality, recent work. 4. Ask AnswerThis to identify a research gap (super useful to new students!). 5. Browse the list of suggested research questions by using the xtract data feature and pick one that feels focused, meaningful, and relevant. 6. Use the Search Papers and Citation Map tools to locate the most relevant studies on your topic. 7. Once you’ve got a strong library of sources, plug your research question into the prompt bar and make sure to specify that you want a thesis statement and add any additional requirements that you would like the thesis statement to meet… 8. And just like that, AnswerThis will generate a full thesis statement backed by credible sources. 9. Use the AI Editor to refine your statement, insert citations in any format, and polish everything to perfection. This means: - No more writer’s block - No more bouncing between tabs - No more guessing if your thesis is solid or what to add Just an AI complete pipeline that helps you clarify, brings confidence, and keeps the momentum going, from the start. Definitely a game changer for all students or researchers, even! Try AnswerThis for yourself. I’ve linked it in the first comment with a 10% discount, thanks to them :) #academicwriting #phdtips #phdstudent #answerthis #aiwritingtools #phdresearch

35 smart ways PhDs can use ChatGPT Most PhDs aren’t fully using ChatGPT. Use it to write faster, think deeper, work smarter. 1) Pick the right model For research or coding, use top-tier models 2) Chain prompts Break big requests into sequenced sub-prompts 3) Role-play Simulate a supervisor, examiner, or peer reviewer 4) Ask for counterarguments Test the strength of your research logic 5) Refine by iteration Revise and rework until it reads perfectly 6) Visual suggestions Generate slide layouts, diagrams, tables, frameworks 7) Generate frameworks Ask for step-by-step plans: methods, lit maps 8) Create a PowerPoint Ask for slide decks for presentations 9) Search GPTs Find useful tools in the GPT marketplace 10) Simulate conversations Practice viva questions or mock interviews 11) Canvas Co-edit thesis, slides, or outlines with collaborators 12) Custom instructions Set tone, depth, and constraints for consistent output 13) Projects Keep chats, notes, and outlines organized by topic 14) Memory Let ChatGPT remember your topic, data, or research area 15) Tasks Reminders for recurring research actions 16) Voice chat Use hands-free mode to brainstorm or rehearse 17) Codex Ask GPT to write or fix code in R, Python, or Stata 18) Socratic prompting Use probing questions to deepen thinking 19) Agent prompting Assign ChatGPT roles like “data analyst” or “thesis coach.” 20) Meta-cognition Ask GPT to critique and improve its output 21) Push back Challenge vague replies for better logic 22) Ask for chapter or section layouts Ask ChatGPT to structure your thesis chapters 23) Use GPT to clarify reviewer feedback Paste comment, ask: “What does this mean?” 24) Check alignment Ask: “Are topic, objectives, and methods connected?” 25) Deep research Ask ChatGPT to go deeper into a topic with references 26) Chat with PDF Upload an article and ask ChatGPT for insights 27) Examiner expectations Ask ChatGPT: “What do examiners expect from this chapter?” 28) Theory suggestions Get best-fit theories with definitions 29) Write in a scholarly tone Ask ChatGPT for help with formal academic writing 30) Check alignment Do your findings answer your questions? 31) Supervisor-style feedback simulation Let ChatGPT act like your supervisor and review your work 32) Journal finder Get journal suggestions with scope, fit, and impact factor 33) Check your contribution Clarify what’s original and why it matters 34) Make GPT explain and improve Ask why it wrote something and refine it 35) Turn on data controls Disable training - keep your work private   Save this post so you don’t forget it! Tag a fellow researcher or PhD student who needs this   Found this helpful? Share & follow for more! -------------------------------------- I help PhD & Masters students simplify research, and complete their thesis with confidence. DM me for expert coaching and support!

💡UX research with AI: Framework for writing effective prompts AI tools are integrated into almost all parts of the design process, including critical ones like user research. User research is integral step in the product design process as it sets the foundation for the entire design process. UXReactor's latest study explores how UX researchers can integrate AI tools like ChatGPT to enhance their processes (https://lnkd.in/d5XzZ26B) The team suggests applying the R.E.F.I.N.E. framework (Role, Expectations, Format, Iterate, Nuance, Example). This framework helps craft more effective prompts, leading to improved outputs in tasks such as creating research plans, screeners, and moderator guides. 🔹 R (Role) Be explicit about who the AI should be. e.g., "Act as a senior UX researcher specializing in healthcare." 🔹 E (Expectations) Set clear goals and outcomes. e.g., "Generate 10 open-ended user interview questions focused on pain points that doctors face." 🔹 F (Format) Specify how you want the output structured. e.g., "List in bullet points with a short rationale for each question." 🔹 I (Iterate) Refine prompts based on outputs. Start broad, then narrow. Provide feedback like: "Make it more concise" or "Focus on usability issues." 🔹 N (Nuance) Include edge cases, audience specifics, or product context. e.g., "Target users are 55+ years old who use assistive tech." 🔹 E (Example) Show a sample of what you expect. e.g., "Here's a format I used before [format], follow this structure." 🔧 Tips to improve your prompts ✔ Pair the framework with real-world project context for best results. ✔ Store prompt templates for repeatable research tasks (screeners, guides, etc.). ✔ Don't rely blindly on output generated by AI; use AI as a collaborator and apply critical thinking. 📕 How to write better prompts for AI design and code generators: https://lnkd.in/dXr8sXBj #UX #research #design #uxdesign #productdesign

ChatGPT o1 and o3 are crazy powerful. But 95% of users don’t know how to unlock their full potential. Let me show you how to unlock their full power: 1. Start with a Precise Research Goal Bad: "Tell me about AI in healthcare." Good: "Give me an in-depth analysis of how AI is transforming disease diagnosis, including current research, ethical concerns, and market trends." The clearer your intent, the better your results. 2. Use Multi-Step Reasoning Instead of a single question, break it down: • Context: What is the background of this topic? • Current Trends: What are the latest advancements? • Challenges: What are the biggest obstacles? • Future: Where is this going in the next 5 years? Example: "Analyze the impact of quantum computing on cryptography. Break it into history, current breakthroughs, risks, and future potential." 3. Force It to Think Like an Expert Use personas to get high-level insights: "Answer this as if you were a cybersecurity analyst writing for a tech conference." "Analyze this from the perspective of an economist forecasting industry growth." You’ll get responses tailored to how experts actually think. 4. Ask for Sources & Contradictions Most people trust whatever AI says but real research needs cross-checking. Prompt: "Cite research papers or case studies on this topic and mention any contradictory viewpoints." This forces it to find diverse and credible perspectives. 5. Refine & Iterate No expert gets it right on the first try. Neither will AI. Follow-up prompts: • "Expand on the ethical concerns in more detail." • "Summarize this in a way a 10-year-old can understand." • "Give me 3 opposing views and their strongest arguments." Each iteration sharpens the research further. BONUS: Plug-and-Play Prompt Template Copy & paste this into ChatGPT (o1 model): "I need an in-depth research report on [TOPIC]. Provide context, current trends, challenges, and future predictions. Use research-backed insights and expert perspectives. Cite sources and mention any contradictions in the field." This will unlock deeper insights than 99% of people ever get. Follow for more

I used to spend 60+ hours searching for research gaps. Now I do it in under 60 minutes. Here's the free AI tool that changed everything. Last year, every new project started the same way: 47 open tabs. 12 half-read papers. Zero clarity. Then a colleague showed me AllSci Corp. It maps 20+ million research papers into hypotheses and evidence. You don't just see what exists. You see what's missing. Last month I typed in my research topic. Within 15 minutes I spotted a connection between two fields nobody had explored. That gap became my latest paper. The tool does 4 things brilliantly: → Organises results by hypothesis, not just keywords → Shows supporting evidence for each claim → Identifies the leading authors you should be reading → Visualises the entire research landscape so you spot gaps others miss Here's how to use it in 4 simple steps: Step 1: Enter your research topic Type your question into the search bar and select "Detailed Search" for AI-powered discovery. Step 2: Review hypothesis-backed results Get results organised by hypothesis and evidence. Skip the noise. Go straight to verified claims. Step 3: Explore hypotheses and key authors Dive deeper into specific hypotheses. Identify the leading researchers you should be reading. Step 4: Visualise the research landscape Explore trends, journals, and connections. Use knowledge graphs to spot gaps nobody else sees. I've mapped out each step with screenshots in the visual below. The best part? It's completely free for researchers. --- What research gap are you currently struggling to define? #ResearchTools #PhDLife

UCSF authors used AI itself to do a systematic literature review, now published as a pre-print review of trial-matching pipelines using large language models. https://lnkd.in/eN8RFvBm This article has a nice secret stated in the open: they acknowledge using Elicit, an LLM-based tool, to perform a systematic literature review. https://elicit.com Lots of home-brewed “literature review” AI projects sprout up, especially in life sciences industry. Sometimes these are IT-types doing a wrapper over document summarization machine that most regular document collection AI systems can now do. These are like NotebookLM, Microsoft’s Copilot Chat Notebook, ChatGPT Projects or Claude’s equivalent, or open source SurfSense. Elicit goes beyond this by helping with steps… steps that most others mess up: 1. Refining the research question, not just taking one and going, and offering one-tap chips to add on refinements. 2. Gathering sources and varying keyword vs semantic search. Because sometimes, like Disney’s Little Mermaid noted, the best articles don’t all have the right… what’s that word again?… keywords. 3. Title and Abstract Screening. Default and user-adjustable criteria are applied automatically and each is scored. This resembles how manual systematic reviews are done. A litmus test for tech-types doing this wrong is they sort of cannonball their way through slipshod check-boxing of articles to include or not. Having criteria that can be replicated is what “systematic” means in “systematic” literature review. 4. Data Extraction. Elicit attempts to extract quantitative and qualitative data, not just summarize, and grab the data presented in tables. This is extraordinarily useful for meta-analysis. 5. Reports with sentence-level citations and formatted with PRISMA-like diagrams and tables as found in real systematic literature reviews. What these systems, homebrewed or not, fail to do is to critique and get at potential holes or missed spots in the original articles — something difficult for humans and often neglected. It requires critical reading and critical thinking skills to realize a scientific publication has missed something, underplayed some data, embellished, or used the wrong method. Similar AI document research tools are emerging in other disciplines. For instance, lawyers reviewing vendor/supplier terms & conditions and privacy policies and contractual agreements can not only rapidly review and propose amendments if needed, especially based on internal corporate legal playbooks, but also also in other legal fields, use tools to rapidly search for and intelligently process case precedent, and legal and regulatory references, and even “red team” their opinions. Those tools might even help legal teams review the terms for those very tools themselves, and also understand if their playbooks are truly mitigating risk… or if they provide only illusory relief (eg if the agreement terms were breached). Cc Samantha Intriligator

👩🏫 Coaching doctoral students through research question (RQ) revision is one of the most rewarding and revealing jobs I have. Recently, I was working with a grad student whose interview-based study explored how high school English teachers experience writing instruction. Her RQs were thoughtful, but one of them risked nudging participants toward a particular kind of answer, a “leading question,” in qualitative terms. We discussed ways to reframe it: Instead of “What training do teachers need...,” try “How do teachers talk about the training that supports...” In that moment, we weren't changing the content of the question but rather exploring different ways to shape its meaning, tone, and implications by expanding her expressive toolkit. 🤖 She asked whether using ChatGPT or Claude might help with that kind of rewording. My response: Absolutely BUT only with a critical mindset. Language models can offer an amazing range of linguistic options. They’re great for brainstorming alternate phrasings or surfacing patterns we might not have seen. Here's what I told her: 🪞 Language models can be valuable tools for expanding our expressive range once we have clarity of communicative intent and conceptual stance. 🪞 Because language is never neutral, these tools can subtly redirect our ideas under the guise of 'wordsmithing.' We have to treat their outputs as ideologically loaded options to be critically sifted. And this is true whether we’re writing a research question, a mission statement, or a LinkedIn post. Words aren’t just words. They’re meaning-making tools. 🧭 Let the thinking lead the language, not the other way around. 🧭

How did we use AI in recent research? With the CGAP project team on harnessing innovation for inclusive finance, we’ve felt it would be a missed opportunity not to use AI. What did we do? The project’s hypothesis was that the regulatory architecture has been evolving in response to the fast-changing financial system. We wanted to understand how exactly and what leaders in the field have done to be more agile and better equipped to address innovation. Key to our research method was literature review, expert interviews, country visits, primary research. We realized the AI could help most with literature review, synthesizing views from a diverse sources on three research questions: 1.     What kind of regulatory responses to innovation are typically associated with better (financial inclusion) outcomes? 2.     What are key enablers and barriers to implementing these responses successfully? 3.     Is there a common set of guiding principles for countries to follow in adoption and implementation of these responses? Here is how we designed the method and the tools we deployed to get results we could use: 1.     Triage. The pool of publications was very diverse as it included laws, regulations, academic papers, and news articles among other sources. We sorted them out into clusters by geographic scope, specific geography, and issuing source (see the figure). 2.     Hierarchy. We have assigned different levels of relevance (hierarchy) to certain publications (e.g., rules vs. guidance vs. research). 3.     Merging. We merged documents in each cluster into one PDF file (or multiple when needed). 4.     Analysis. Using ChatPDF and simple contextual prompts, we have asked the research questions for each cluster. Once done, we did the meta-analysis synthesizing responses related to all clusters. 5.     Audit trails. We recorded each step in the process in a way that would allow us to review it later, replicate it and/or try different prompts. 6.     Human in the loop. As the team (Denise Dias, Stefan Staschen, Mehmet Kerse, and me) kept working in parallel on interviews, country visits, and limited literature review, we articulated our own responses to the research questions that we could then compare with the AI results. Then, Bryce F., in charge of the AI research, showed us the results. They were mixed for the first two questions, possibly distorted by the noise we could have filtered out with better triage. But the answer to our third question was largely spot on. This exercise has helped us quickly, with minimum effort identify the areas where our own findings resonated with the prevalent opinions, highlight the areas where our insights diverged (requiring extra validation) or added new perspectives. The exercise saved us hours of scanning through the materials. We could have used ChatGPT (or similar) to finetune the method. We could have used an AI agent to help with triage. Maybe next time. What’s your experience with AI in research?