The Short Answer
To research for a debate, work in four passes: a survey pass to map the topic, a depth pass to read the strongest arguments on each side, a evidence-cutting pass where you pull verbatim quotes with full citations, and a refutation pass where you draft responses to the cards you just collected. Most novice debaters do only the first pass and walk into round with opinions instead of evidence.
The research itself isn't the hard part. The hard part is structuring what you find so you can retrieve it under time pressure, defend the source on cross-examination, and update it when the topic evolves. This guide covers the full workflow used in policy, public forum, and Lincoln-Douglas debate, plus a faster version for classroom and informal debates.
Why Research Wins Rounds
In any judged debate, the team with better evidence has a structural advantage that good speaking can't fully overcome. Judges, especially flow judges, weight specific cited evidence higher than general analysis. A 2024 study from the National Speech and Debate Association found that in elimination rounds, the side citing more recent and topic-specific evidence won 68% of the time, controlling for speaker rank.
But research doesn't just give you evidence to cite. It gives you the conceptual map of the topic — the actual disagreements between experts, the strongest arguments on each side, and the empirical claims that are contested versus settled. Walking into round with that map is the difference between debating the resolution and debating your prejudices.
For the prep work that comes after research, see how to prepare for a debate.
Pass One: The Survey Pass
The survey pass takes one to three hours depending on topic complexity. The goal is not to find evidence yet — it's to understand the landscape.
Start with overview sources
Read three to five overview pieces before reading anything else. Good sources for the survey pass:
After the survey pass, you should be able to write a one-paragraph summary of the topic, list the three to five major positions, and name the empirical questions that the debate turns on. If you can't, read more.
Map the disagreement
Open a document. Write the resolution at the top. Underneath, write three columns: "Affirmative arguments," "Negative arguments," and "Contested empirical claims." As you survey, fill in the columns. By the end, you should have five to ten arguments per side and a list of factual disputes that determine who's right.
This document becomes the table of contents for the rest of your research. Every later source you read maps onto something in these columns.
Pass Two: The Depth Pass
Now you read the strongest arguments on each side, slowly, with notes.
Find the canonical sources
Every debate topic has a small number of canonical sources — the books, papers, and reports that everyone in the field cites. You need to know what they say. To find them:
For a typical resolution, you'll end up with eight to twelve canonical sources. Read all of them. Yes, this takes time. Debaters who skip this step lose to debaters who don't.
Read against your position
Spend more time reading the side you disagree with than the side you agree with. Two reasons. First, your refutations need to come from a real understanding of the opposing case, not a strawman. Second, the strongest arguments on the side you'll be assigned to defend in negation rounds will come from sources you initially disagreed with.
When you encounter an argument that surprises you or makes you uncomfortable, mark it. That's a flag — either you've found something the opposing side will hit you with, or you've found a contention you should adopt yourself.
Take structured notes
For each source, write:
These notes become the spine of your case file. Skipping them feels efficient and isn't.
Pass Three: The Evidence-Cutting Pass
Now you turn raw research into usable evidence. In policy debate this is called "cutting cards"; in public forum it's "building blocks"; in academic research it's "extracting quotations." The mechanics are the same.
What makes a good card
A good card has four properties:
Cards that fail any of these are weaker than they look. Stack ranking your cards by these criteria during the cutting pass saves you in round when you have to choose which one to read.
How to format a card
Industry standard format for competitive debate:
\\\
TAG: One-sentence summary of what the card proves.
CITATION: Author Last Name, "Title," Publication, Date. URL or stable identifier.
QUOTED EVIDENCE: The verbatim quote, with the most relevant
sentence underlined or highlighted. No ellipses inside the
underlined section.
\\\
The tag is the part you'll read aloud first. It should be the conclusion you want the judge to draw. The citation is what you'll read second. The evidence is what you'll read third, and only the underlined portion needs to be read in time-constrained formats.
For the speech mechanics around how cards fit into a constructive, see how to write a debate speech.
Vet every source
For each card you cut, ask:
Sources that fail vetting either get dropped or get noted as weak so you don't read them when a stronger card exists.
Pass Four: The Refutation Pass
This is the pass that separates serious debaters from casual ones. For every card you've cut for your own side, you draft the refutation your opponent will use, then draft your response.
The threat model
For each card, write down:
Most attacks on evidence come in three forms: the source is biased, the methodology is flawed, or the conclusion doesn't follow. Each requires a different defense, prepared in advance.
Build a blocks file
A "blocks file" is a document where each common argument gets a pre-written response. Structure it so each block is two to four sentences plus a citation. In round, you find the relevant block and read it; you don't compose from scratch.
A typical blocks file for a topic has fifty to one hundred blocks. Starting from zero feels overwhelming. Most debaters share blocks files within a school or partnership, building incrementally each season.
For the underlying refutation logic, see how to refute an argument.
Tools That Help
The actual tools matter less than the workflow, but here's what works for most debaters:
Don't let tool selection delay you. The best debaters use whatever they already have well.
Research Ethics in Debate
Three rules that come up in every debate handbook and that judges will sanction violations of:
These rules apply equally in academic, policy, and informal debate. The cost of breaking them is asymmetric — a single fabricated card can end a debate career.
A Faster Workflow for Classroom Debates
If you have one to two days rather than weeks, compress the four passes:
Total: under three hours. Not enough for a national tournament, plenty for a classroom round. The discipline of the four passes matters more than the time spent in each.
For students new to the format, start with debate for beginners before scaling up to full prep.
Common Mistakes That Waste Research Time
Reading without note-taking. Two hours of reading without notes is one hour of reading you'll have to redo. Take notes.
Cutting cards before understanding the topic. Cards cut without context tend to be off-topic or vulnerable. Survey first, cut second.
Over-collecting evidence. A round uses ten to twenty cards, not two hundred. Past a certain point, more cards make retrieval slower without making rounds easier to win.
Ignoring methodology. Empirical claims rest on methodology. If you don't understand how a study was conducted, you can't defend the card on cross-examination.
Researching only your side. Half your prep should be opposition prep. Otherwise you'll lose negation rounds reliably.
Frequently Asked Questions
How long should research take for a tournament topic? Twenty to forty hours for a competitive policy or LD topic, distributed across the season. Public forum topics turn over monthly, so eight to fifteen hours per topic is typical.
Where can I find old debate evidence files? Many circuits share evidence files openly through OpenEvidence, NDCA Open Caselist, and similar repositories. Use these as starting points, not finished products — adapt the cards to your case and verify each source.
What if my topic is so new there isn't much research? Read primary sources directly — the actual policy, the actual ruling, the actual study. New topics reward debaters who go to primary sources because secondary commentary hasn't formed yet.
How do I avoid plagiarism if I'm using shared cards? Plagiarism in debate has a specific meaning: presenting someone else's research as your own to a judge. Citing the original author of evidence is not plagiarism. Reading another team's analysis as if it were your own is.
Do AI tools help with research? For survey-pass overviews and locating sources, yes. For cutting cards verbatim, no — AI hallucinates citations and quotes, which would constitute fabrication. Always verify primary sources yourself.
How does research connect to the case I bring into the round? Research feeds the brief; the brief feeds the case; the case is what you read from in the round. The most efficient workflow researches both sides simultaneously, then drafts the case using the strongest 3-5 pieces of evidence per contention rather than dumping everything found. For the dedicated case-writing workflow with a working template, see how to write a debate case.
Once I have the evidence, how do I actually use it in a speech? Research finds evidence; integration determines whether judges weight it. The Warrant + Evidence + Impact pattern — state the warrant in your own words first, cite the source briefly, then state the impact in terms the judge can weigh — is what separates evidence that scores from evidence that decorates. For the integration patterns, citation conventions across formats, and the comparative-evidence move that wins close rounds, see how to use evidence in a debate.
Ready to put your research to work? Practice debating against AI on Debate Ladder.