Christopher I. Rider

Adapted from Chapter 8 of the book manuscript Yeah, Yeah, But… and Therefore, and from an earlier handout (PDF).

Your paper is an argument. Your abstract is the “elevator pitch” for that argument, and your introduction is the “opening statement” to the jury. Both must be direct, logical, and relentlessly focused on demonstrating your contribution.

Write the abstract first, and let it reject ideas

Most researchers write the abstract last, as a faithful precis of a completed paper. This guide asks you to do the opposite: write it first, and write it aspirationally, suspending disbelief and stating every claim as if it were already fully supported. This is not dishonesty; the aspirational abstract is clearly labeled, to yourself, as the version of the paper you wish you had. It is the target, not the draft. And its purpose is not to describe a project but to decide whether the project deserves your time.

The arithmetic of a research career makes this discipline non-negotiable. Good ideas vastly outnumber the ones a person can see through to publication; over a career the ratio of ideas entertained to papers published runs roughly ten to one. The scarce resource is not ideas but the months of attention each finished paper consumes. An abstract written first, and read honestly, lets an idea die on the page in an afternoon rather than dying in a journal’s rejection after two years. Letting ideas die cheaply is not a failure of the method; it is the point of it.

So the rule is: do not write the introduction until the abstract excites you, and do not continue the paper until the introduction does. If you cannot draft six sentences that would make you want to read the paper, the problem is not your writing but that the contribution is not yet there, and no amount of analysis will install one that the frame cannot find.

The abstract: six sentences, six jobs

An abstract that contributes does six jobs, in order, one to a sentence. The sequence is not a template to be filled but the same arc the rest of the book has been walking (affirm what the reader believes, introduce what unsettles it, reconcile the two), compressed to its smallest legible form. The unsettling is not optional ornament. A claim earns a reader’s attention, as Davis (1971) argued, only when it denies an assumption the audience held; affirm everything and you have written a summary, not a contribution. This is the same “deny a prior” move Chapter 1 asked of your question, now carried into the sentence that announces what the project found.

Your abstract is a short, logical argument, not a summary. Each sentence must perform a specific job.

• Sentence 1: The Hook. Start with the big puzzle, paradox, or unresolved debate that your paper addresses. This establishes why the topic is important and “joins the conversation.”
  • Heuristic (The “So What?” Test): Can a colleague from a different field read this first sentence and understand why this is a problem worth solving?
• Sentence 2: The “Gap” or “Problem.” Explain why the puzzle from Sentence 1 is still unresolved. This is the “However...” moment. You’re identifying the “usual suspect” explanations and arguing they are “insufficient,” “incomplete,” or “immeasurable”.
  • Heuristic (The “Important Gap” Test): Are you just “gap spotting” (which Barney, 2020, calls “unacceptable”), or have you explained why this gap has important consequences that need to be addressed?
• Sentence 3: Your Core Contribution. Introduce your novel mechanism, framework, or argument as the solution. This is the “Aha!” moment of the abstract. Give your concept a clear name (e.g., “valuative bias”, “socially endogenous calibration”, “push vs. pull”).
  • Heuristic (The “Aha!” Test): Is this the single, novel idea you want the reader to remember and associate with your paper?
• Sentence 4: The Empirical Test. State where and how you test your argument. This is your “ideal laboratory”. Justify the context by its unique advantages (e.g., “Analyzing over 1,300 NFL coaches...”, “Using... individual-level prestige ratings...”). This grounds your theory in rigorous evidence.
  • Heuristic (The “Strategic Context” Test): Does this sentence make it clear why your data or setting is uniquely suited to isolating your mechanism, solving a problem that has “limited” prior work?
• Sentence 5: The Key Finding. State your answer to the puzzle from Sentence 1, in the language of the mechanism from Sentence 3. Be direct and unambiguous.
  • Heuristic (The “Payoff” Test): Does this sentence directly answer the puzzle you opened with? (e.g., “We find that White assistant coaches were promoted at higher rates...”; “We find stronger evidence of access than for selection or development...”).
• Sentence 6: The Implication (The “So What?”). Conclude by stating the broad takeaway. How does your finding change how we think about the “big conversation” from Sentence 1? This is the “payoff” that makes your paper important.
  • Heuristic (The “New Agenda” Test): Does this sentence point to a new research agenda, a new policy implication, or a new strategic prescription?

A worked example

The mentorship example uses simulated data, generated for illustrative purposes.

The six-sentence abstract, with each sentence doing one job. Read sentences 1–3 as the “Yeah, yeah…,” sentence 4 as the design that earns the “…but…,” and sentences 5–6 as the “…and therefore.” The worked example is the mentorship project carried through the book; the simulated truth behind sentence 5 is +8 pp.

Three failures recur often enough to name. The first is a Sentence 3 that names two mechanisms instead of one. The contribution sentence must give the reader a single clean label to carry away; the rival accounts all belong in the diagram, which lives in the theory section, not the abstract. If sentence 3 hedges between sponsorship and selection, the reader leaves with neither. The second is a Sentence 4 that describes the data rather than the design: “one million observations from 1998 to 2026” tells the reader the size of your dataset and nothing about why it can separate your stories. Sentence 4 names the strategic context that does the identifying: the lottery, the discontinuity, the shock. The third is a Sentence 5 that apologizes. The abstract is the home of the finding, not the caveat, and the standard is calibrated force: state the finding without apology, bounded by what the design can actually carry (an unhedged sentence that is nonetheless true to the design’s reach).

For a published abstract that does all six jobs, Syverson (2004) is a model of the form. In a short paragraph it puts the phenomenon on the table (persistent productivity dispersion within narrow industries), names what the dominant technological account leaves unexplained, commits to one mechanism (demand-side spatial substitutability), describes the strategic context that identifies it (geographic variation created by the industry’s high transport costs), reports the finding (denser markets show a higher productivity floor and mean and less dispersion), and draws the implication (market structure, not only technology, shapes the productivity distribution). Read it sentence by sentence and the six jobs are visible; read it whole and it is the replicate–diverge–reconcile arc compressed to a paragraph.

From abstract to introduction: the opening statement

The introduction is the same argument as the abstract, expanded, and the right model for it is an attorney’s opening statement to a jury. Here is the premise; here is the evidence I will present; here is the belief you hold now; here is the belief I want you to hold by the end. You write it the way you wrote the abstract, sentence by sentence, taking each abstract sentence as a headline and saying more because now you have the room. Aim for two to three pages, and spend the extra room developing the arc (the affirmations, the pivot, the reconciliation), not on adding caveats the abstract was careful to omit.

Six paragraphs, six jobs, mirroring the six sentences:

• Hook. Open with the puzzle, not a literature review. State the phenomenon and the question it forces in a way that makes a reader want the next paragraph. The reader should not have to ask why they are reading this.
• “Yeah, yeah.” Affirm the established explanation generously. Here is what the field already knows, stated as its proponents would state it. You are loading the prior you intend to move; an unaffirmed reader has nothing for your pivot to act on.
• “But.” Name the anomalous fact the established account cannot explain, the specific ignorance the paper will reduce. Not a gap for its own sake, but the precise point the comfortable story gets wrong.
• Contribution. Name and label the mechanism, and contrast it with the prior. This is where the Elster diagram belongs: introduce the rival stories on equal footing and commit to the one your design will defend. Tell the reader, here, what would distinguish it from the alternatives.
• Empirical strategy. And not before the fifth paragraph. Lead with the data and you have asked the reader to evaluate your design before they care about your question. By now they care; describe the strategic context that identifies the mechanism, briefly; the table does the rest.
• Roadmap. The structure of what follows, and what the reader will see. This paragraph is also a private checklist: it names, out loud, the empirical burdens the rest of the paper has just promised to discharge.

One rule governs the whole sequence. A contribution must extend, refine, replace, or reconcile an existing understanding (Barney, 2020); merely noting that no one has studied something (gap-spotting) is not a contribution, because an unexamined corner is not yet a belief anyone holds, and you cannot move a belief no one has. The same objection is pressed at length by Zuckerman (2008, 2015), who declines to count “the literature has overlooked” as a puzzle framing at all: a phenomenon earns attention when it makes trouble for a theory, not when it is merely unstudied. The mentorship introduction does not say “no one has randomized mentoring.” It says “the field believes mentoring causes promotions; we show most of that belief is selection.” The first is a hole. The second is a reconciliation, and only the second earns the reader’s time.

References