psychbull-open-science-and-transparency

Open Science & Transparency (psychbull-open-science-and-transparency)

Effective February 1, 2022, Psychological Bulletin requires transparent reporting for empirical work, including meta-analyses, aligned with the TOP Guidelines. Authors who submit quantitative analyses must supply their database, codebook, and relevant scripts, and must state whether the design/hypotheses were preregistered and where to access them. Build this as you go — it is not a last-minute add-on. (Verify the exact TOP level wording on the live page; see 待核实 in the source map.)

When to trigger

• Before screening — to preregister the protocol
• Assembling the data/code/materials package for submission
• Answering the journal's transparency/availability statements
• A reviewer asks where the database, codebook, or scripts live

Preregister the protocol (before screening)

• Register the review protocol on PROSPERO (the systematic-review registry) or OSF Registries: question, eligibility criteria, search plan, coding plan, and the pre-specified analysis and moderators.
• The manuscript must state whether work was preregistered and where to access it.
• Masking note: Bulletin submissions registered on PROSPERO may attach registration details as an appendix, omitting author names and the registration number so review stays masked.
• Clearly distinguish pre-specified (confirmatory) from exploratory analyses in the writeup.

The transparency package (TOP)

1. Database — the coded effect-size dataset (one row per effect, with study identifiers and moderator codes), in an open format.
2. Codebook — variable definitions and decision rules (from psychbull-inclusion-and-coding).
3. Analysis scripts — code that reproduces every pooled estimate, moderator model, bias diagnostic, table, and figure.
4. Materials — search strings/logs, screening records, and the PRISMA-flow inputs.
5. Deposit to a trusted repository (OSF / institutional) with a persistent identifier (DOI); complete the journal's data-availability statement, explaining any restricted data.

Anti-patterns

• Preregistering after screening (defeats the purpose)
• "Available upon request" instead of a deposited, identifiable package
• A dataset no one else could re-run (no codebook, undocumented variables)
• Leaving the PROSPERO number / author names in an appendix during masked review
• Blurring confirmatory and exploratory analyses

What transparency referees verify

As an APA flagship that adopted TOP-aligned reporting, Psychological Bulletin treats the transparency package as a reviewable deliverable, not a courtesy. The bar reviewers apply:

Transparency item	Pass	Reject / desk-return trigger
Protocol timing	Preregistered before screening, location stated	Registered after coding, or "available on request"
Database deposited	One-row-per-effect file with a DOI	No data, or a spreadsheet no one else can run
Codebook deposited	Variable definitions and decision rules included	Undocumented columns; codes unexplained
Scripts deposited	Reproduce every estimate, table, and figure	Numbers in text that no script regenerates
Masking preserved	PROSPERO appendix stripped of number and names	Registration number visible during masked review

Worked vignette — assembling the package

Illustrative figures only. For the self-affirmation synthesis (k = 42, g = 0.34), the deposit under this skill's rules:

• Protocol on OSF Registries timestamped before screening, listing eligibility, the search plan, and the delivery-format moderator as confirmatory.
• Database: 51 rows (42 studies, 9 contributing multiple effects), each with study ID, g, variance, and moderator codes, in an open CSV with a DOI.
• Codebook carried over from the inclusion-and-coding stage, defining all 23 variables.
• Scripts: one metafor/robumeta script regenerates the pooled 0.34, the meta-regression (R²-analog 0.22), every bias diagnostic, the forest and funnel plots, and the summary table.
• Masking: the PROSPERO/OSF appendix is anonymized, with the registration number and author names removed, so review stays masked.

Referee pushback → venue-specific fix

• "Was this preregistered, and where?" → Add the registry link and the timestamp predating screening; state explicitly which analyses were confirmatory.
• "'Data available on request' is insufficient." → Deposit the database, codebook, and scripts to a trusted repository with a persistent DOI and complete the availability statement.
• "Confirmatory and exploratory analyses are blurred." → Separate them in both the manuscript and the deposited scripts so the distinction is auditable. (Confirm the exact TOP level on the live page.)

Output format

【Protocol】preregistered (PROSPERO/OSF) before screening? [Y/N] + where
【Masking】PROSPERO appendix anonymized (no number/names)? [Y/N]
【Database + codebook】deposited with DOI? [Y/N]
【Scripts】reproduce all estimates/exhibits? [Y/N]
【Availability statement】drafted; restricted data explained? [Y/N]
【Confirmatory vs exploratory】clearly separated? [Y/N]
【Next】psychbull-submission

Supplementary resources

• ../../resources/external_tools.md — PROSPERO, OSF, TOP, repositories, reproducibility tooling
• ../../resources/official-source-map.md — TOP (Feb 1 2022), data/codebook/scripts, preregistration/PROSPERO policy