L’interface est localisée ; les présentations et documents des Skills sont traduits automatiquement par IA dans votre langue. Le code, les commandes, chemins et termes techniques restent en version originale, et le document de l’auteur fait foi.
Explorer les Skills
Données et recherche · SKILL.md

devpsych-submission

devpsych-submission

Use when running the final pre-submission preflight for Developmental Psychology (APA) via Editorial Manager — length-tier compliance, the abstract and Public Significance Statement, APA 7th + JARS reporting, masked-review anonymization, the data-availability statement, and preregistration/data DOIs. Final checks; it does not draft content.

Independent GitHub snapshot · Automated classification · Original documentation governs

Documentation de la Skill

GitHub

name: devpsych-submission description: Use when running the final pre-submission preflight for Developmental Psychology (APA) via Editorial Manager — length-tier compliance, the abstract and Public Significance Statement, APA 7th + JARS reporting, masked-review anonymization, the data-availability statement, and preregistration/data DOIs. Final checks; it does not draft content.

Submission Preflight (devpsych-submission)

The last check before submitting through Editorial Manager (editorialmanager.com/dvl). Three things sink Developmental Psychology submissions most often: the developmental claim not being credible (age vs. cohort, invariance, attrition), JARS/length-tier non-compliance, and weak transparency. Verify volatile specifics (length numbers, abstract ceiling, policy wording, editor) on the official page first.

When to trigger

  • "Submitting tomorrow" — last pass before upload
  • Unsure what the length tier, abstract, and transparency requirements demand
  • Confirming the Public Significance Statement, anonymization, and deposits are in order

Process facts (verify volatile items on the official page)

  • Publisher: American Psychological Association (APA). ISSN: 0012-1649 (print); 1939-0599 (online).
  • Editor: Koraly Pérez-Edgar (Penn State; editorship since 2023-01-01) (web-verified 2026-06-22; re-verify before submission).
  • Portal: Editorial Manager, editorialmanager.com/dvl.
  • Review: masked peer review available on request — keep author identity out of the manuscript and out of shared links; submit a masked copy via a stable link or supplemental material.
  • Length tiers (检索于 2026-06;以官网为准): ~4,500 words (brief report), ~10,500 (larger report), ~15,000 (multi-study / comprehensive longitudinal); length counts cover page, abstract, text, references, tables, and figures.
  • Front matter: APA-style abstract plus a Public Significance Statement (2–3 sentences, ~30–70 words) on the abstract/keywords page.
  • Style: APA 7th edition; JARS reporting (JARS-Quant / JARS-Qual / MARS).
  • Transparency: TOP guidelines — data-availability statement, persistent identifiers (DOIs), preregistration where applicable, sample-size justification.

Preflight checklist

Format & front matter

  • Manuscript within the chosen length tier (total counts cover page, abstract, text, refs, exhibits)
  • APA-style abstract within the current ceiling; developmental change foregrounded
  • Public Significance Statement present: 2–3 plain sentences (~30–70 words)
  • APA 7th edition style; tables/figures and references formatted per APA

Developmental credibility (JARS)

  • Age vs. cohort addressed; change claim matched to the design
  • Measurement invariance reported before interpreting change (configural→metric→scalar)
  • Attrition analysis + principled missing-data model (FIML/MI)
  • Sample-size justification for the change parameter (power/precision)
  • Effect sizes + confidence intervals for major results; full JARS disclosure

Masked review (if requested)

  • No author names/affiliations in the masked manuscript
  • Repository / preregistration links anonymized (e.g., OSF anonymized view; Databrary view link)
  • Identifying file metadata stripped

Transparency (TOP)

  • Data-availability statement present (what is shared / restricted, why)
  • Data + scripts deposited with DOIs (or a justified restricted-access archive); reproduce in a fresh session
  • Minors'/identifiable data handled per consent/IRB (e.g., permission-based archive)
  • Preregistration linked; confirmatory/exploratory consistent

Worked micro-example — preflight verdict (illustrative)

Running the three-wave effortful-control package through the gate the night before upload:

Tier:       larger research report; total 9,800 / ~10,500 words → PASS
Abstract:   within ceiling; foregrounds within-person growth → PASS
Public Sig: 3 plain sentences, 58 words → PASS
Credibility: scalar invariance reported; FIML + attrition analysis;
            power simulation for the interaction → PASS
Stats:      slope and interaction carry effect sizes + 95% CIs → PASS
Mask:       no names; OSF anonymized-view + Databrary view links;
            PDF metadata stripped → PASS
Transparency: data-availability statement present; data/scripts DOI'd;
            video via Databrary; prereg linked → PASS
Verdict:    ready to upload to Editorial Manager (editorialmanager.com/dvl).

Last-mile failure modes that trigger a return

FailureCaught byQuick fix
Over the length tierformat checkmove robustness/full models to online supplemental material
No Public Significance Statementfront-matter checkadd 2–3 plain sentences (~30–70 words)
Trajectory with no invariance testcredibility checkreport configural→metric→scalar before change
Listwise deletion, real attritioncredibility checkrefit with FIML/MI; add attrition analysis
OSF/Databrary link reveals authormasking checkswitch to anonymized/view links before upload
"Data on request," no statementtransparency checkadd a data-availability statement + DOIs

Volatile specifics (length numbers, abstract ceiling, accepted article types, current editor) change. Re-confirm every number above against the journal's current submission guidelines before relying on it.

Anti-patterns

  • Submitting over the length tier instead of using online supplemental material
  • Missing or jargon-laden Public Significance Statement
  • Interpreting change with no invariance test, or deleting dropouts
  • Non-anonymized manuscript or links during masked review
  • No data-availability statement / transient links instead of DOIs

Output format

【Tier】within the chosen length limit? [Y/N]
【Front matter】abstract + Public Significance Statement (2–3 sentences)? [Y/N]
【Credibility】age/cohort + invariance + attrition + sample-size justification? [Y/N]
【APA 7th + JARS】style + ES/CI + full disclosure? [Y/N]
【Masked】manuscript + links + metadata anonymized? [Y/N]
【Transparency】data-availability statement + DOIs + preregistration? [Y/N]
【Next】await decision → devpsych-rebuttal on R&R

Supplementary resources

Skills enregistrées