Expecon Writing Style

Writing Style (expecon-writing-style)

When to trigger

• The abstract names a topic but never states the treatment contrast or the headline effect
• A reader reaches the results before understanding what subjects actually did
• The design is described as scattered implementation notes instead of a clean procedure section
• The intro reads like a JEBO behavioral essay or a GEB theory paper rather than an experimental-methods paper

Structure: the design is the narrative

ExpEcon papers have a near-canonical, design-forward shape. Honor it; referees read for it.

1. Abstract (the contract). Within a few sentences: the question, the treatment contrast, the incentivized design in one clause, the headline effect with a number and direction, and the takeaway. "We run a lab experiment in which [contrast]; [outcome] is X% higher under [T] (p=…); this adjudicates between [model A] and [model B]."
2. Introduction. Motivate, state the precise hypothesis, then say how the design isolates it and what is new about that design. The reader should know your treatments before the literature review ends.
3. Design / Procedures (the load-bearing section). This is what an experimentalist reads most carefully. Write it so the study is reproducible from the text alone: game and payoffs (ECU + conversion), treatments and what each manipulates, matching, information, rounds, incentive mechanism, subject pool and recruitment, software (z-Tree/oTree), sessions and n, comprehension checks. State explicitly that no deception was used.
4. Predictions. The signed, treatment-indexed hypotheses from expecon-theory-model.
5. Results. Raw behavior → primary test (correct unit) → secondary/exploratory (labeled) → robustness.
6. Discussion. Mechanism, external validity, boundary conditions, and what the design cannot speak to.

Voice an ExpEcon referee trusts

• Procedural transparency over rhetoric. The register is sober and precise. Claims are pinned to the design ("because matching was anonymous and one-shot, reputation cannot drive this"), not to adjectives.
• Subjects' perspective. Describe what participants saw and decided, in their units; do not narrate only the analyst's regression.
• Say "no deception" out loud. Experimentalist readers look for the explicit statement that participants were not deceived and were paid real, salient incentives. Make it unmissable.
• Calibrate the claim to control. Internal validity is your strength; do not overclaim external validity. A line acknowledging the population (often a student pool) and what generalization is and is not warranted reads as competence, not weakness.
• Hedge honestly on nulls. If the result is a precise null, frame it with the power to detect a meaningful effect; that is a strength here, not an apology.

Sentence-level habits

• Name treatments, do not number them only. "the Punishment treatment" reads far better than "T3"; if you must abbreviate, define mnemonic labels (NoPun / Pun) and use them consistently.
• Quantify in the subjects' units. "contributions rose by 3.2 tokens (16% of the 20-token endowment)" beats "contributions rose significantly."
• Attribute causality to the design, not the data. "Because matching was anonymous and one-shot, reputation cannot explain X" is the ExpEcon move; "X is correlated with Y" is not.
• Front-load the procedure of each elicitation before its result, so the reader knows what was incentivized.
• Use the past tense for what subjects did and the present for what the data show — a small convention experimentalist readers expect.