journal_venue_selection v1.0.0

name: "journal_venue_selection" description: "Choose where to submit a paper before anyone adapts it for a venue: build a ranked shortlist of preprint servers and journals from the authors' own priorities (speed, reach, cost, review time, fit for the work), apply hard gates first, run a scoring sheet with the authors' weights, check each venue's current rules on AI assistance and disclosure with the deciding rule quoted from the live policy page, check legitimacy with Think Check Submit and DOAJ, and write down why each venue made or missed the list. Use when authors ask where to submit, where to preprint, which journal fits this paper, or whether a venue is legitimate, in any field. Includes a runnable scoring script and a worked shortlist for an education research paper." metadata: { "includeInPrompt": true }

Journal Venue Selection

Purpose

Decide where a paper goes, in what order, and why, before anyone rewrites the paper for a venue. The output is a ranked shortlist the authors can defend: their priorities turned into weights, hard gates applied before any scoring, every surviving venue checked against its live scope, fee, preprint and AI policies, and a written reason for each venue that made the list and each one that missed it.

This skill is field neutral. It works for a journal article, a preprint, or a paper that needs both, in any discipline. It does not adapt a manuscript to a venue, write a cover letter, or judge whether the science is good. It answers one question well: given what these authors care about, where should this work go first, second, and not at all.

Two failure modes this skill exists to stop. The first is choosing by journal metric or habit, then discovering a fee the authors cannot pay, a scope rule that rules the paper out, or an AI disclosure rule the draft already breaks. The second is submitting to a venue nobody checked, including venues that charge fees without real peer review.

Workflow

1. Pin the paper and the authors' priorities before naming any venue. Record: the paper's field, method, and one sentence claim; who the work must reach; and the authors' priorities in their own order. The usual five are fit for the work, reach, cost, speed, and review time or review quality. Ask for a hard budget for article fees, any funder or institution open access mandate, any deadline that is real (a tenure clock, a conference, a thesis), and whether the authors plan to preprint. If the authors cannot rank their priorities, propose an order from what they said and get them to confirm it. Weights invented by the agent and never confirmed are the most common way a shortlist goes wrong.

1. Turn the priorities into hard gates and weights. A hard gate is a pass or fail rule that no score can outvote: a mandatory fee above the stated budget with no waiver or no fee route, a scope statement that rules this paper type out, a funder mandate the venue cannot meet, or a journal preprint ban when the authors will preprint. Everything else becomes a scored criterion, 1 to 5, with weights that follow the confirmed priority order. Write the gates and weights down before scoring anything. Changing a weight after seeing the ranking is allowed only if the authors change a priority, and the change gets recorded with its reason.

1. Build a long list from the field, not from memory of famous names. Sources that work across fields: the journals the paper already cites, the journal lists of the relevant scholarly society, the index or database the field actually uses, and the preprint server the field's community uses. For preprint servers, start from the server list for the discipline: OSF hosts field servers such as EdArXiv for education research, PsyArXiv for the psychological sciences, and SocArXiv for the social sciences, and other fields have their own established servers. Aim for six to ten candidates across journals and at most two preprint servers. Cut at the long list stage without scoring when there is no recorded scope fit. Scoring everything is how a weak venue sneaks in on one strong column.

1. Run the legitimacy check on every journal candidate. Use the Think. Check. Submit. journal checklist and, for open access journals, a direct lookup in the Directory of Open Access Journals (DOAJ). The checks that decide most cases: you or colleagues know the journal and can find its latest papers; the publisher is easy to identify and contact; the journal is clear about the type of peer review it uses and does not guarantee acceptance or a very short review time; articles are indexed or archived in services the field uses, verified on the index site itself and never taken from the journal's own claim; fees are clear about what they charge, what for, and when; the editorial board is recognisable and named with affiliations; the publisher belongs to a recognised body such as COPE, DOAJ, or OASPA. A venue that fails two or more of these is gated out, and the failed checks are the recorded reason. Listing in DOAJ is positive evidence. Absence from DOAJ is a reason to check harder, not a verdict on its own.

1. Check fit against the live scope, and quote the deciding line. For each surviving journal, read the current aims and scope and the author guidelines, then find recent papers like this one in the journal's own archive. Record the one scope sentence that decides fit, quoted, with the page and the check date. Scope pages often carry the warning that matters most, such as a journal ruling out small single site studies or papers where a tool is only a delivery platform. Score fit from that evidence, not from the journal title. If no deciding sentence can be found, fit stays unscored and the venue stays off the ranked list until it can.

1. Check the preprint pairing before anything goes public. Decide the preprint server and the first choice journal together. Read the journal's policy on prior posting and cross check it in Sherpa Romeo, which collects publisher preprint and self archiving policies. Most journals now allow a prior preprint. Some still treat one as prior publication, and that pairing is a hard gate. Then check the server's own gates: field fit, moderation, and any submission limits or endorsement rules. As one dated example, arXiv changed its endorsement policy on 21 January 2026 so an institutional email alone no longer qualifies a new submitter. Server rules move, so check the live help pages and record the date.

1. Run the AI policy check pass on every surviving venue, with the deciding rule quoted. This step is mandatory, and references/policy-quotes.md holds the verified baseline quotes. For each venue, open the publisher or journal policy page that is live today and record four things: whether AI tools may be listed as an author (COPE and ICMJE both say no, because an AI tool cannot take responsibility for the work), what AI use must be disclosed, where the disclosure goes, and what is exempt. Placement is where authors get caught, because publishers differ. Elsevier wants a standalone declaration at the end of the manuscript before the reference list, in its own format, and exempts basic grammar, spelling and punctuation checks. Nature Portfolio wants LLM use documented in the Methods section or a suitable alternative part, and exempts AI assisted copy editing under a narrow definition. Quote the deciding rule for the actual venue, name the page, and date the check. If the authors used AI help of any kind, map their actual use against each venue's rule and flag any venue whose rule the current draft would already break, for example AI generated images at a publisher that does not permit them. A venue whose AI rule the authors cannot or will not meet is gated out, with the quote as the reason.

1. Record cost, speed and review facts from the venue, dated, or leave them blank. Article fees, waiver routes, submission to first decision times, and review models come from the journal's own pages or its publisher journal insights page, recorded with the check date and marked volatile. Never carry a fee or timeline from a third party ranking site, a colleague's memory, or an earlier shortlist. If a fact cannot be confirmed from the venue in this pass, the row stays blank, the criterion stays unscored for that venue, and the shortlist says so in plain words. An unconfirmed fee is the classic case: the venue goes on the reserve list, not the ranked list.

1. Score with the script, then read the result like an editor. Put the sheet in the format of templates/venue-sheet.csv: one row per venue, a hard_gate column, one column per criterion, and a notes column that carries the evidence for the scores. Run python3 scripts/venue_score.py <sheet.csv> --weights fit=3,reach=2,speed=1,cost=2,review=1, using the authors' confirmed weights and their criterion names. The script returns weighted totals on a 1 to 5 scale for venues that pass their gates, and lists gated out venues separately with the scores they would have had, so the cost of a gate stays visible. Then sanity check the ranking by hand: if the top venue is one the authors would not actually send the paper to, a criterion or a gate is missing, and that gets fixed in the sheet, not by hand editing the ranking.

1. Write the shortlist as a decision record. Follow the shape of references/worked-shortlist.md. For each venue: rank, kind, total, the deciding scope quote, the AI rule quote with its placement, the dated cost and speed facts or the plain statement that they are unconfirmed, and one line on why it made the list. For each cut venue: the gate or check it failed, quoted or named, in one line. End with the submission sequence: preprint step first where one is planned, then journal order, and what triggers a move to the next venue (a decline, a desk rejection, or a deadline passing). Date the whole record. A shortlist older than the policies it quotes is a draft, not a decision.

Output Contract

• A ranked shortlist, usually two to four journals plus at most one chosen preprint server, each with its weighted total and the weights that produced it.
• For every ranked venue: the deciding scope quote, the AI disclosure rule quoted with where the disclosure goes, and dated fee and timing facts, or an explicit note that a fact is unconfirmed and what that blocks.
• A gated out and cut list, one line per venue, naming the gate, the failed legitimacy check, or the missing evidence that removed it.
• A reserve list for venues that fit but carry an unconfirmed fact, with the exact check that would bring them back.
• The submission sequence, including the preprint step and its journal policy check, and the trigger for moving to the next venue.
• The scoring sheet CSV and the script output, kept with the shortlist so the ranking can be rerun when priorities change.
• No venue presented as suitable on the strength of a journal metric alone, and no fee, timeline, acceptance rate, or policy stated without its source page and check date.

Operating Rules

1. Every factual claim about a venue comes from that venue's or publisher's live pages, an index checked at the source, or a standards body, verified in this run. Fees, timelines, indexing, and policies are volatile. Date every one, and cut or flag anything that cannot be verified rather than smoothing it over.
2. The authors' priorities set the weights. Never substitute a prestige ranking, a journal metric, or your own preference for their order. The San Francisco Declaration on Research Assessment (DORA) general recommendation applies here in plain form: do not use journal based metrics such as Journal Impact Factors as a surrogate measure of the quality of an individual paper.
3. Hard gates run before scores, always. A venue that fails a gate is listed as gated out with its would be score visible, never quietly dropped and never promoted by a high total.
4. Quote the deciding rule for AI assistance and disclosure for each venue from the live policy page. Publisher baselines from COPE, ICMJE, Elsevier, or Springer Nature orient the check. They do not replace the venue's own current instructions, and a disclosure written for one publisher does not satisfy another.
5. Check the journal's preprint policy before recommending a preprint, and check the preprint server's own submission rules in the same pass. Keep any editor reply that clarifies an unclear policy with the shortlist.
6. Treat solicitation email from an unknown journal as a lead to check, never as evidence of suitability. Run it through the same legitimacy check, and expect most of it to fail.
7. Scores are judgements, and the shortlist says so. Facts get sources and dates. Judgements get reasons. Never dress a judgement up as a measured fact, and never present an acceptance rate or review time that the venue itself does not publish.
8. Style for anything written with this skill: plain sentences, no em dashes, no promotional language about any venue. A venue is a fit for this paper and these authors, or it is not. It is never described in sales language.