Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› psci-theory-and-hypotheses
psci-theory-and-hypotheses
GitHub用于撰写心理科学论文的理论与假设部分，强调在数据前声明理论及确认性假设。区分预注册与探索性分析，避免HARKing，明确证伪条件及适用范围，符合期刊透明度规范。
Trigger Scenarios
撰写理论框架和假设
准备预注册或预分析计划
调和预测与实际发现
审稿人指出存在事后假设或无理论依据
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill psci-theory-and-hypotheses -g -y
SKILL.md
Frontmatter
{
"name": "psci-theory-and-hypotheses",
"description": "Use when stating the theory and hypotheses for a Psychological Science manuscript. The journal's credibility norms require a clear separation of confirmatory (preregistered) from exploratory analyses and theory stated before results. Structures the argument; it does not run analyses."
}
Theory & Hypotheses (psci-theory-and-hypotheses)
Psychological Science rewards a clear theoretical claim and honest hypothesis status. Because preregistration quality is weighed and a Research Transparency Statement is required, the cardinal rule is: state theory and confirmatory hypotheses before the data, and label exploratory results as exploratory.
When to trigger
- Writing the theoretical framing and hypotheses
- Preparing a preregistration / pre-analysis plan (Registered Report or OSF)
- Reconciling what you predicted with what you found
- A reviewer flagged the work as "post hoc," "atheoretical," or "HARKed"
Build the argument
- State the theory. What mechanism or principle predicts the effect, and why. Keep it crisp — the format does not allow a long theoretical treatise.
- Derive directional hypotheses. Translate the theory into specific, testable predictions (direction and, where possible, magnitude/effect-size expectations).
- Mark hypothesis status. Separate confirmatory (preregistered, predicted in advance) from exploratory (generated after seeing data) analyses — clearly, in the text.
- Specify what would disconfirm. Say which results would count against the theory; this is what makes a test credible.
- Match to the right format. If the theory is best tested prospectively and confirmatorily, a
Registered Report makes the claim far stronger (route via
psci-review-process).
Avoiding HARKing (Hypothesizing After Results are Known)
- Do not present an exploratory finding as if it were predicted.
- If you discovered something post hoc, say so and frame it as a hypothesis for future confirmation — Psychological Science values this honesty and the Transparency Statement surfaces it.
Worked micro-example — theory to hypothesis status (illustrative)
The two-study attention package, written so hypothesis status is legible at a glance.
Theory: Brief mindfulness induction transiently broadens attentional
control, reducing automatic capture by emotional distractors.
H1 (confirmatory, preregistered, S1+S2):
Induction reduces capture cost vs. control (directional; d ≥ 0.30 SESOI).
H2 (exploratory in S1 → confirmatory in S2):
Effect is larger for high trait-anxiety participants
(interaction; preregistered only for Study 2).
Disconfirming: a CI for H1 centered near zero, or a reversed sign, counts
against the broadening account — stated up front.
Constraints-on-generality (a venue expectation)
Post-credibility-revolution, name the boundary of the claim explicitly rather than implying universal scope. Confirm current wording against the journal's submission guidelines, but include the substance:
Constraints on generality: We expect H1 to hold for healthy adults under
brief laboratory inductions and screen-based capture tasks. We do not claim
it generalizes to clinical anxiety, sustained meditation training, or
real-world attentional demands without further test.
Theory-stage reviewer pushback and the venue fix
| Reviewer pushback | Psychological Science fix |
|---|---|
| "Atheoretical / mechanism unclear" | state the broadening mechanism in one crisp sentence before the hypotheses |
| "This looks HARKed" | show the preregistration timestamp; relabel post hoc analyses exploratory |
| "Claims outrun the manipulation" | scale H1's scope to what the induction plausibly does; add a constraints-on-generality clause |
| "Non-directional hypothesis" | commit to a sign and, where possible, a SESOI magnitude |
Theory calibration anchors
- Theory here is crisp, not encyclopedic: one mechanism sentence and a derived, directional prediction beat a multi-page treatise that the 2,000-word budget cannot hold.
- The cardinal honesty rule is temporal: anything specified before data is confirmatory; anything generated after seeing data is exploratory, full stop. The Research Transparency Statement and a graded preregistration make the distinction visible to editors and reviewers.
- Stating what would disconfirm the theory is what converts a story into a test; reviewers steeped in the credibility revolution read an unfalsifiable framing as a red flag.
- When the theory is best tested prospectively and confirmatorily, a Registered Report makes the claim
materially stronger — confirm current Registered Report availability against the journal's submission
guidelines, then route via
psci-review-process.
Anti-patterns
- Theory written to fit the result after the fact (HARKing)
- Vague, non-directional "we expect a relationship" hypotheses
- No statement of what would disconfirm the theory
- Blurring confirmatory and exploratory analyses
- A theory section so long it breaks the 2,000-word budget
Output format
【Theory】the mechanism/principle, briefly
【Hypotheses】directional, testable (H1, H2, …)
【Status】which are confirmatory (preregistered) vs exploratory
【Disconfirming evidence】what would count against the theory
【Format fit】Research Article vs Registered Report (if confirmatory + prospective)
【Next】psci-study-design
Supplementary resources
../../resources/external_tools.md— preregistration templates and tools../../resources/official-source-map.md— preregistration and transparency policy
Version History
- 1839142 Current 2026-07-05 14:15
