Expecon Referee Strategy

Referee Strategy (expecon-referee-strategy)

When to trigger

• The paper is analytically done and you want to pre-empt the objections an ExpEcon referee will raise
• You suspect a specific weak point (a borderline procedure, modest power, a confound) and want to disarm it before submission
• You need to decide what to add now vs. defend in the letter later
• You want to read the draft the way two editors and two anonymous referees will

Who reviews you here, and what they look for

Experimental Economics uses anonymous (double-blind) refereeing, and accepted papers need the approval of two editors (检索于 2026-06；以官网为准). Your referees are practicing experimentalists. They are unimpressed by topical novelty and intensely focused on design integrity. They read in a predictable order: gates first, then the contrast, then inference, then reproducibility. Pre-empt each.

The standard pushbacks, ranked by lethality

1. "This is deception / the incentives aren't real." Lethal and binary. If any procedure could be read as deceptive, name it explicitly and explain why it is not deception under the ESA definition (withholding ≠ misstating). Confirm every elicitation is incentive-compatible and report realized payments. Do not leave the referee to wonder.
2. "It's underpowered / the unit of inference is wrong." The most common substantive rejection. Pre-empt with a power/MDE statement at the session/matching-group level and group-level inference. A precise null beats a noisy positive.
3. "A confound, not the mechanism." Referees hunt for the alternative explanation: comprehension differences, session effects, order/learning, demand effects, experimenter effects. For each plausible alternative, show the design rules it out or report the test that does.
4. "The treatment contrast is impure." If a treatment changes more than one thing, a referee will say the effect is unattributable. Show the minimal-pair logic, or add the control treatment that isolates the dimension.
5. "The behavioral models aren't actually distinguished." If rival accounts predict the same data, the experiment is undiscriminating. Show where predictions diverge and that your design sits on that divergence.
6. "Not reproducible / instructions missing." Have the instructions and z-Tree/oTree code ready at submission; referees check them for deception and comprehension.
7. "Why here and not JEBO/GEB?" If a referee thinks the contribution is topical (JEBO) or theoretical (GEB), they may recommend transfer. Make the methods contribution explicit.

Turning anticipated objections into pre-emptive moves

• For each top objection, decide: fix now (add a treatment/analysis), disclose and bound (report the limit honestly with the test), or defend in text (a paragraph that pre-rebuts). Build a small internal table before submitting.
• Add a short "threats to validity" or "robustness" subsection that names the alternatives and the test for each — referees reward authors who do their job for them.
• Suggesting appropriate (non-conflicted) referees and clearly framing the methods contribution in the cover letter can shape who reviews and how.

A pre-submission self-review pass

Read your own draft in the referee's order and answer out loud:

1. Gates — Could any sentence in Procedures be read as deception? Are all incentives real and incentive-compatible? Is this stated explicitly?
2. Power — What is the MDE at the group level, and did the study have power to detect it? Would a referee call the n of groups (not subjects) thin?
3. Confounds — List every alternative explanation a smart experimentalist would propose; for each, is there a design feature or test that kills it?
4. Purity — Does each treatment manipulate exactly one thing?
5. Discrimination — Do my contrasts actually separate the competing models, or do they all predict the same pattern?
6. Repro — If a referee opened the appendix right now, are the instructions and code there and legible?

If any answer is weak, that is the item to fix or pre-empt before the desk editor ever sends it out.

The demand-effect and experimenter-effect family

Beyond the headline gates, experimentalist referees reliably probe a cluster of subtle validity threats — treat them as first-order, not nitpicks:

• Demand effects — did the design telegraph the hypothesis? Pre-empt with neutral framing, a post-experiment guess question, and (if needed) an obfuscated cover task.
• Experimenter effects — could the experimenter's presence or scripting bias responses? Use standardized on-screen instructions and double-blind payment where social-image is at stake.
• Order/learning effects — counterbalance and test, or justify a between-subject design.
• Selection into the lab — acknowledge the recruited pool and what it does and does not represent.
• House money / wealth effects — pay one random round; do not let accumulated earnings contaminate later choices.

For each, the winning move is the same: name the threat in a short validity subsection and show the design feature or test that addresses it, so the referee sees you raised it first.