The key, published in fullDer Schlüssel, vollständig veröffentlicht
Last post walked through how I rebuilt the AQ-50’s scoring key — item by item, from face validity, flagging the three calls I wasn’t fully confident on. That post made the case for publishing the actual key rather than just asserting a score was computed from one. This is that: the full table, not a summary of it.
A quick reminder of the shape of it, since it’s easy to lose in 50 rows: each item is answered on a 4-point scale (definitely/slightly agree, slightly/definitely disagree), but only scores 0 or 1 — a point only if the response lands on whichever side is keyed below. 26 items are keyed so disagreeing is the point-scoring direction; the other 24 are keyed so agreeing is. Total range 0–50.
The condensed version
Disagree scores a point (26 items): 1, 3, 8, 10, 11, 14, 15, 17, 24*, 25, 27, 28, 29*, 30, 31, 32, 34, 36, 37, 38, 40, 44, 47, 48, 49*, 50
Agree scores a point (24 items, everything else): 2, 4, 5, 6, 7, 9, 12, 13, 16, 18, 19, 20, 21, 22, 23, 26, 33, 35, 39, 41, 42, 43, 45, 46
* = items 24, 29, 49 — the three I flagged as logic-inferred rather than confidently recalled. See the last post for why.
The full table, with reasoning
One row per item: which answer scores the point, and the one-line reasoning behind that call. Deliberately not the original statement wording itself (not mine to republish) — just the key and the logic, same as I'd reconstruct it from scratch.
| # | Key | Why |
|---|---|---|
| 1 | Disagree | Preferring solitary activity over doing things with others is the trait direction. |
| 2 | Agree | Preferring sameness/routine is a core trait. |
| 3 | Disagree | Difficulty visualizing/imagining is trait-consistent; "very easy" is the non-trait answer. |
| 4 | Agree | Intense, narrow absorption to the exclusion of surroundings is a trait. |
| 5 | Agree | Heightened sensory detail-detection (sound) is a trait. |
| 6 | Agree | Fascination with detail-strings (plates, codes) is a trait. |
| 7 | Agree | Misjudging social appropriateness of one's own speech is a trait. |
| 8 | Disagree | Difficulty picturing story characters reflects weak imagination/theory-of-mind. |
| 9 | Agree | Narrow fascination with a category (dates) is a trait. |
| 10 | Disagree | Difficulty tracking multiple simultaneous social threads is a trait. |
| 11 | Disagree | Social difficulty is the core trait being measured. |
| 12 | Agree | Noticing details others miss is a trait. |
| 13 | Agree | Preferring quiet/solitary space over a social gathering is a trait. |
| 14 | Disagree | Difficulty inventing stories reflects weak imagination. |
| 15 | Disagree | Being drawn to objects/systems over people is a trait. |
| 16 | Agree | Intense circumscribed interests, distressing when blocked, is a trait. |
| 17 | Disagree | Disliking small talk is a trait. |
| 18 | Agree | Difficulty with conversational turn-taking (monopolizing) is a trait. |
| 19 | Agree | Narrow systemizing fascination (numbers) is a trait. |
| 20 | Agree | Difficulty inferring a character's intentions reflects weak theory-of-mind. |
| 21 | Agree | Disinterest in fiction sits in the same imagination subscale as #3/#8/#14. |
| 22 | Agree | Difficulty forming new friendships is a trait. |
| 23 | Agree | Constant pattern-detection/systemizing is a trait. |
| 24* | Disagree | Preferring a systematic/factual venue (museum) over a social-performance one (theater) is the trait direction — agreeing with "theater" is the non-trait answer. |
| 25 | Disagree | Being upset by routine disruption is the trait; disagreeing with "does not upset me" means it does. |
| 26 | Agree | Difficulty sustaining a conversation is a trait. |
| 27 | Disagree | Difficulty with inferential/implicit language ("reading between the lines") is a trait. |
| 28 | Disagree | Detail-focus over gestalt/whole-picture processing is the trait; disagreeing with "whole picture" means detail-focus. |
| 29* | Disagree | Being good at remembering number-strings fits the systemizing/detail-memory pattern (same family as #6/#19); disagreeing with "not very good" means good at it. |
| 30 | Disagree | Noticing small changes is a trait; disagreeing with "don't usually notice" means noticing. |
| 31 | Disagree | Difficulty reading nonverbal social cues (boredom) is a trait. |
| 32 | Disagree | Difficulty multitasking/attention-switching is a trait. |
| 33 | Agree | Difficulty with phone-call turn-taking is a trait. |
| 34 | Disagree | Preferring planned over spontaneous activity is a trait. |
| 35 | Agree | Difficulty with humor/implicit meaning (missing the point of a joke) is a trait. |
| 36 | Disagree | Difficulty reading facial expressions is a trait. |
| 37 | Disagree | Difficulty resuming a task smoothly after interruption (attention-switching) is a trait. |
| 38 | Disagree | Difficulty with small talk is a trait. |
| 39 | Agree | Perseverating on a narrow topic, noticeably to others, is a trait. |
| 40 | Disagree | Difficulty with/dislike of pretend play as a child is a trait. |
| 41 | Agree | Systemizing/categorical collecting behavior is a trait. |
| 42 | Agree | Difficulty with perspective-taking is a trait. |
| 43 | Agree | Preferring careful planning over spontaneity is a trait. |
| 44 | Disagree | Disliking social occasions is a trait. |
| 45 | Agree | Difficulty inferring others' intentions is a trait. |
| 46 | Agree | Anxiety/resistance toward novel situations is a trait. |
| 47 | Disagree | Disliking meeting new people is a trait. |
| 48 | Disagree | Lacking social tact/diplomacy is a trait. |
| 49* | Disagree | Same logic as #29 — good at remembering date-of-birth facts fits the pattern; disagreeing with "not very good" means good at it. |
| 50 | Disagree | Difficulty with pretend play (same family as #40) is a trait. |
Table shown once, not duplicated per language — item numbers and the agree/disagree calls read the same either way, and retranslating fifty rows of reasoning risked introducing new errors into a table whose whole point is being checkable.
Why bother publishing the table at all
The condensed version above is everything you need to actually score a run. The full table exists for the opposite reason: not to use, but to argue with. If item 28’s reasoning looks wrong to you, or item 7’s, that’s the table doing its job — a bare “here’s the key” asks for trust, a reasoned one invites a correction. Three of these I’ve already flagged as shaky myself; there may be more I didn’t catch.
Der letzte Beitrag ging durch, wie ich den Auswertungsschlüssel des AQ-50 neu aufgebaut habe — Aussage für Aussage, aus der Plausibilität heraus, mit einer Markierung der drei Entscheidungen, bei denen ich nicht voll sicher war. Dieser Beitrag hat dafür argumentiert, den tatsächlichen Schlüssel zu veröffentlichen, statt nur zu behaupten, ein Wert sei daraus berechnet worden. Das hier ist das: die vollständige Tabelle, keine Zusammenfassung davon.
Kurz zur Erinnerung, wie das Ganze aufgebaut ist, da es sich in 50 Zeilen leicht verliert: Jede Aussage wird auf einer 4-Punkte-Skala beantwortet (stimme definitiv/eher zu, stimme eher/definitiv nicht zu), zählt aber nur 0 oder 1 — ein Punkt nur, wenn die Antwort auf die unten festgelegte Seite fällt. 26 Aussagen sind so geschlüsselt, dass Ablehnung die punktende Richtung ist; die anderen 24 so, dass Zustimmung es ist. Gesamtbereich 0–50.
Die kompakte Version
Ablehnung zählt einen Punkt (26 Aussagen): 1, 3, 8, 10, 11, 14, 15, 17, 24*, 25, 27, 28, 29*, 30, 31, 32, 34, 36, 37, 38, 40, 44, 47, 48, 49*, 50
Zustimmung zählt einen Punkt (24 Aussagen, alles andere): 2, 4, 5, 6, 7, 9, 12, 13, 16, 18, 19, 20, 21, 22, 23, 26, 33, 35, 39, 41, 42, 43, 45, 46
* = Aussagen 24, 29, 49 — die drei, die ich als logisch hergeleitet statt sicher erinnert markiert habe. Siehe den letzten Beitrag für das Warum.
Die vollständige Tabelle, mit Begründung
Eine Zeile pro Aussage: welche Antwort den Punkt gibt, und die kurze Begründung dahinter. Bewusst nicht der originale Aussagentext selbst (den darf ich nicht einfach weiterveröffentlichen) — nur der Schlüssel und die Logik, so wie ich ihn von Grund auf neu aufbauen würde.
Tabelle oben nur einmal gezeigt, nicht pro Sprache dupliziert — Itemnummern und die Zustimmen/Ablehnen-Entscheidungen lesen sich in beiden Sprachen gleich, und fünfzig Zeilen Begründung neu zu übersetzen hätte riskiert, neue Fehler in eine Tabelle einzubauen, deren ganzer Sinn Nachprüfbarkeit ist.
Warum die Tabelle überhaupt veröffentlichen
Die kompakte Version oben ist alles, was man braucht, um tatsächlich einen Durchgang auszuwerten. Die vollständige Tabelle existiert aus dem gegenteiligen Grund: nicht zum Benutzen, sondern zum Widersprechen. Wenn die Begründung von Aussage 28 für dich falsch aussieht, oder die von Aussage 7, dann macht die Tabelle genau das, wofür sie da ist — ein bloßes „hier ist der Schlüssel“ verlangt Vertrauen, ein begründeter lädt zur Korrektur ein. Drei davon habe ich selbst schon als wackelig markiert; es könnten weitere sein, die ich nicht bemerkt habe.