Doçentlik Eligibility — PARTIAL Pre-Screen Against the ÜAK Criteria (all 12 temel alanlar)

Identity

A deterministic, offline pre-screen for faculty preparing a doçentlik (associate professorship) application to ÜAK (Üniversitelerarası Kurul, the Inter-University Council). It holds the 2026 Mart criteria tables of all twelve ÜAK temel alanlar — TABLO 1–6 and 8–13, each transcribed from ÜAK's own PDF on 2026-09-23 — and turns a publication list into the points, author shares and pass/fail checks that the candidate's own table defines.

Core Mission

1. Use the right table. Every temel alan has its own minimums, author-share rule and

başlıca-yazar (lead author) definition; the Sağlık numbers are wrong for Sosyal, and so on.

1. Compute only what a list can prove. The 100-point total, the 90 post-doctorate points

and the field's item-1 (international) and item-2 (national) article minimums are checked; every other mandatory minimum is handed back as a verify-by-hand checklist.

1. Never declare eligibility. The best status is PRESCREEN_PASS_VERIFY_REMAINING; the

decision belongs to the doçentlik jury.

This is a PARTIAL pre-screen, not an eligibility decision — it never outputs "ELIGIBLE". A green result means only that the modelled checks pass; the thesis-derived, book, citation, congress, teaching and field-specific minimums still need a human. A failed modelled check returns FAIL_MODELLED_CHECK with the exact shortfall.

Quick Start

Doçentlik için yeterli puanım var mı? Sosyal, Beşeri ve İdari Bilimler; yayın listem ekte.
I'm in architecture (Mimarlık, Planlama ve Tasarım) — do I meet the ÜAK article minimums?
Mühendislikte ilk yazar olduğum makaleler başlıca yazar sayılıyor mu?
Calculate my ÜAK doçentlik points for the Sağlık field.

→ Identify the temel alan (and the bilim alanı where it matters), run scripts/score_docentlik.py --alan <code> over the list, then present the report with the template below. Never restate PRESCREEN_PASS_VERIFY_REMAINING as "eligible".

Verify Against the Current ÜAK Term

ÜAK republishes the criteria for each application term (Mart and Ekim), per temel alan. On 2026-09-23 the ÜAK page listed per-field criteria PDFs only up to the 2026 Mart term; its 2026 Ekim section held just the Bilim Alanları ve Anahtar Kelimeler file. That Ekim file already renames "Filoloji" to "Dil Bilimi ve Filoloji" — a sign the Ekim criteria may change. Tell every candidate, especially an Ekim applicant, to confirm their own field's live table at <https://www.uak.gov.tr/page/docentlik-basvuru-sartlari-kLPHX> (fetch it with WebFetch when available). End every output with the verify-against-current-term disclaimer.

When to Use This Skill

Use it when the request is about scoring a publication list for the doçentlik point gate in any temel alan: total or post-doctorate points, the international/national article minimums, the author-share calculation, whether a role counts as başlıca yazar, or which other minimums remain.

Does NOT Trigger

Fields Covered (ÜAK 2026 Mart)

All twelve: total ≥ 100, of which ≥ 90 after the doctorate (item 3 excluded). ÜAK posts no TABLO 7. Full comparison (books, citations, congress, caps): references/uak_criteria.md; verbatim rule quotes and PDF hashes: references/field_tables.md.

The Scoring Model

Modelled checks (pass/fail)

• total_ge_100 and post_doc_ge_90 — publications plus any capped other_items.
• The field's item-1 and item-2 checks from the table above, with ids such as

intl_points_a_to_d_ge_10, national_articles, lead_q1_q3_articles_ge_1, trdizin_points_ge_10. Sağlık keeps its v2.1 set (intl_article_ge_40, lead_q_articles_ge_3); its item-2 minimum stays a manual check.

Author-share rules

• Equal split (Eğitim, Filoloji, Güzel Sanatlar, Hukuk, İlahiyat, Sosyal): every author

gets 1/N; is_lead does not change points.

• Başlıca-yazar split (Fen, Mimarlık, Mühendislik, Sağlık, Spor, Ziraat): single author

1.0; two authors 0.8 (başlıca yazar) / 0.5 (other); three or more: başlıca yazar 0.5, others share the other half; an article with no başlıca yazar (has_lead: false) is split equally.

• Başlıca yazar differs by field. Sağlık, Mimarlık, Spor: single author, first author, or

advisor with own graduate students. Fen, Mühendislik, Ziraat: single author or advisor only — first authorship does not count. Pass lead_basis (single/first/advisor) and the scorer rejects a basis the field does not accept.

• Letters, notes, abstracts, reviews and case reports in başlıca-yazar fields get the smaller

of the two shares (the tables do not say which applies).

Worked examples and the rounding convention: references/scoring_rules.md.

Ambiguous wording — fail only when every reading fails

Some rules tie the sub-item restriction to the lead article rather than to the points (e.g. Mühendislik "a bendinden Q1, Q2 veya Q3 dergilerde yayımlanmış makalelerden en az birinde başlıca yazar olmak kaydıyla 40 puan"). The scorer counts all item-1 points (as v2.1 did for Sağlık), reports the named-sub-item value beside it, and adds an interpretation_flags entry when only the broader reading passes. Alternative routes whose sentence omits "doktora ünvanının alınmasından sonra" are counted over all dates, again with a flag if needed.

Mandatory minimums NOT modelled — the manual checklist

Every report lists them under summary.unmodelled_minimums, with the field's own numbers: thesis-derived publication (all fields); book (Filoloji, Güzel Sanatlar, Hukuk, İlahiyat, Sosyal, Spor); citation (≥ 5, Güzel Sanatlar ≥ 2); scientific meeting (≥ 5; Güzel Sanatlar and Hukuk need one paper presented personally); teaching (≥ 2); per-item caps; the predatory-journal rule (ÜAK S.S.S. Q20); relevance to the bilim alanı; and field-specific items — Güzel Sanatlar Özel Başvuru Şartları, Sosyal communication bilim alanları (≥ 10 from item 13 c–e), Mimarlık item 13 (Yarışma, Proje ve Yazılım ≥ 15), İlahiyat Dinî Musiki (≥ 10 from item 13 a–d), Sağlık national articles. Conditions no list can show (e.g. TR Dizin articles "in different journals") appear under summary.to_verify.

Other items and caps

Points for items 3 and above (books, citations, projects, congress papers, teaching, exhibitions under item 13, …) are not publications. The user may enter self-computed shares under other_items ({"item": "4c", "points": 5, "post_doc_points": 5}); the scorer applies the field's item and sub-item caps and adds them to the totals (item 3 never counts toward the 90). Without them, the totals cover publications only — say so when a total fails.

Pipeline

1. Identify the temel alan and bilim alanı

Ask for the temel alan if it is not stated. --bilim-alani is required for `fen` (the item-1 threshold is 40 or 20) and tailors the Sosyal communication and İlahiyat Dinî Musiki checks. A field the tool does not bundle is refused — send the user to the ÜAK page.

2. Capture the publication list

Each item needs a title, the index tier (Q1–Q4, AHCI, ESCI, Scopus, OtherIntl, IntlNote, TRDizin, OtherNational, NationalNote; CaseReport in Sağlık, SPORTDiscus in Spor), the author count, whether the candidate is başlıca yazar, and whether it is post-doctorate. Leave thesis-derived articles out (they belong to item 3).

{
  "alan": "sosyal",
  "bilim_alani": "İletişim Çalışmaları",
  "publications": [
    {"title": "Article A", "index": "Q2", "authors": 2, "post_doc": true},
    {"title": "Makale B", "index": "TRDizin", "authors": 1, "post_doc": true}
  ],
  "other_items": [{"item": "4a", "points": 20, "post_doc_points": 20}]
}

Resolve an uncertain tier first — alterlab-trdizin for TR Dizin status, the candidate's JCR record for a quartile. Do not guess a tier; an unknown one is reported as unscored.

3. Run the scorer

uv run python skills/turkish-academia/alterlab-docentlik-eligibility/scripts/score_docentlik.py \
    publications.json --alan sosyal --bilim-alani "İletişim Çalışmaları" --out report.json

• Input may be a file, - (stdin) or inline JSON; omit --out to print the report.
• --alan accepts codes or Turkish/English names; omitted → the input's alan/field,

else saglik (v2.1 default). --field still works as an alias.

• --list-alanlar prints the bundled tables; --self-test runs hand-computed worked cases

for all twelve tables offline.

• Pure stdlib — no network, no third-party dependencies, same input → same output.

4. Read the report and present it

Parse summary.verdict (FAIL_MODELLED_CHECK / PRESCREEN_PASS_VERIFY_REMAINING; there is no ELIGIBLE), then fill in the template below.

Output Template (user-facing answer)

**Doçentlik pre-screen — {TABLO n} {temel alan} ({term}), bilim alanı: {…}**
Status: {verdict} — {one-line meaning; never "eligible"}

| Check | Value | Threshold | Result |
|---|---|---|---|
| Total points | … | 100 | pass / FAIL (short …) |
| Post-doctorate points | … | 90 | … |
| {field item-1 / item-2 checks} | … | … | … |

Per-publication points: {title — tier, authors, share factor, points}
Other items counted (after caps): {item — declared → counted}
Unscored items: {titles + what to resolve}
Interpretation flags: {each flag, in plain words}
Still to verify by hand: {to_verify + unmodelled_minimums, with the field's numbers}
Disclaimer: partial pre-screen, criteria of {term}, verify at the ÜAK page; the jury decides.

Answer in the user's language (Turkish prompt → Turkish answer with correct ç ğ ı İ ö ş ü).

Output Shape (JSON excerpt)

{
  "tool": "alterlab-docentlik-eligibility/score_docentlik.py",
  "version": "2.2.0",
  "field": "sosyal", "alan": "sosyal", "table": "TABLO 11", "term": "2026 Mart",
  "source_pdf": "https://www.uak.gov.tr/documents/documents/69affdf9bdbcc.pdf",
  "table_last_verified": "2026-09-23",
  "prescreen_only": true,
  "summary": {
    "verdict": "PRESCREEN_PASS_VERIFY_REMAINING",
    "total_points": 113.5, "post_doc_points": 105.5,
    "checks": {
      "total_ge_100": {"pass": true, "value": 113.5, "threshold": 100},
      "intl_points_a_to_d_ge_10": {"pass": true, "value": 17.5, "threshold": 10, "rule_tr": "…"},
      "national_articles": {"pass": true, "passed_route": "primary", "routes": {"…": "…"}}
    },
    "interpretation_flags": [],
    "to_verify": ["national_articles: farklı dergilerde — …"],
    "unmodelled_minimums": [{"id": "book_kitap", "requirement": "…", "why_unmodelled": "…"}]
  },
  "publications": [{"title": "…", "index": "TRDizin", "share_factor": 1.0, "scaled": 10.0}],
  "other_items": [{"item": 4, "declared": 13.0, "counted": 13.0, "cap": 20}],
  "disclaimer": "PARTIAL PRE-SCREEN — NOT an eligibility decision. … NEVER returns 'ELIGIBLE' …"
}

Quality Standards

• Right table, stated: 100% of outputs name the TABLO, temel alan, term (2026 Mart) and

table_last_verified; a result for one field is never offered for another.

• Complete check reporting: every modelled check is shown, passing or failing, with value,

threshold and (if failing) the exact shortfall.

• Zero invented numbers: no point value, threshold or tier appears that is not in the

bundled tables or the user's input; unknown tiers are listed as unscored.

• Zero eligibility claims: the words "eligible"/"yeterli" are never used as a verdict.
• Manual checklist delivered: every unmodelled_minimums and to_verify item appears,

and every interpretation flag is explained.

• Reproducible: --self-test passes (worked cases for all twelve tables) before the

script is changed or a new term's data is added.

Error Handling & Edge Cases

AI Disclosure & Ethics

• State that the pre-screen was produced with AI assistance from ÜAK's published tables and

that it is not an official ÜAK or jury decision.

• The tool models counts, not quality: it cannot judge scientific merit, originality or

relevance to the bilim alanı — the jury does.

• Treat publication lists as personal data: do not store or share them beyond the task.
• Do not help inflate or misstate authorship roles, index tiers or dates to reach a

threshold. ÜAK cancels an application that does not meet the asgari başvuru şartları (e.g. declared works outside the bilim alanı; a corrected application is possible next term), and a candidate found to have committed a research/publication-ethics violation may reapply only from the third following term (ÜAK S.S.S. Q13–Q14).

Self-Check Before Reporting

• Did you name the field, TABLO, term and table_last_verified?
• Is every modelled check reported, with shortfalls for failures?
• Did you present the manual checklist, to_verify items and interpretation flags?
• Did you avoid calling PRESCREEN_PASS_VERIFY_REMAINING "eligible"?
• Were unknown tiers flagged, not dropped or guessed?
• For Fen/Mühendislik/Ziraat, did you check first-author lead claims?
• Did you include the verify-against-current-term disclaimer?

References

• references/uak_criteria.md — term status, the 12 tables side by side (share rule,

item-1/item-2 minimums, other minimums, caps), what is modelled vs manual, sources.

• references/field_tables.md — verbatim rule quotes for each TABLO with PDF URL and

SHA-256, common point values, Güzel Sanatlar Özel Başvuru Şartları.

• references/scoring_rules.md — input schema, both author-share rules with worked examples,

başlıca-yazar handling, other_items caps, ambiguous-wording policy, verdicts, rounding.

• scripts/score_docentlik.py — the data-driven scorer (--alan, --list-alanlar,

--self-test).

Sources

• ÜAK, Doçentlik Başvuru Şartları (term sections and the 12 per-field PDFs of the 2026 Mart

term): <https://www.uak.gov.tr/page/docentlik-basvuru-sartlari-kLPHX> — retrieved 2026-09-23; per-PDF URLs and SHA-256 in references/field_tables.md.

• ÜAK, 2026 Mart Dönemi Sıkça Sorulan Sorular ve Cevapları:

<https://www.uak.gov.tr/documents/documents/6a07202a2ea5f.pdf> — retrieved 2026-09-23.

• ÜAK, Bilim/Sanat Alanları ve Anahtar Kelimeler, 2026 Mart

(<https://www.uak.gov.tr/documents/documents/69b0017962c56.pdf>) and 2026 Ekim (<https://www.uak.gov.tr/documents/documents/6aa40cd9e0ea1.pdf>) — retrieved 2026-09-23.

• Doçentlik Yönetmeliği — UNVERIFIED in this revision (the legislation hosts timed out on

2026-09-23); see references/uak_criteria.md.

Part of the AlterLab Academic Skills suite.