scoring guide

Practice feedback only; current ETS rubrics informed these original paraphrased summaries. Read the complete official writing/speaking scoring guides in sources/ for full descriptor detail. Do not treat a model response as a unique answer or guaranteed 5. Judge an actual response holistically, not by counting required phrases. Objective teaching positions 107:Q1–50 Reading,Q51–97 Listening,Q98–107 sentence building. One point for each correct response. MC letters must match the selected content. Letter-gap answers supply the exact missing letters; capitalization is irrelevant except where it changes the word. Complete words can receive learning feedback but do not follow the missing-letter direction. Sentence building preserves supplied chunk boundaries, fixed text and grammatical meaning; sentence-initial capitalization is allowed. Do not award the point merely for using the right vocabulary in an ungrammatical order. Evaluate any defensible alternative order against the actual chunk inventory before rejecting it. There is no partial objective score. Write an Email (Q108)0–5:5 clear, effective, developed purpose with consistently suitable social conventions and varied precise language;4 mostly effective with adequate support and few errors;3 generally completes the task, with partial elaboration or noticeable language/register problems;2 limited or unclear message and accumulated errors;1 very little development, disconnected language, serious errors, or minimal original wording;0 blank, non-English, rejected/unrelated topic, entirely copied prompt, or arbitrary keystrokes. Q108 task anchors: changed recorder availability; group member's departure and effect on Thursday recording; a feasible alternative; identifiable help or approval requested from the coordinator. These guide relevance, not a four-box automatic score. Greeting/closing alone cannot compensate for an ineffective message. Academic Discussion (Q109)0–5:5 relevant, clear, well-developed contribution with varied precise language;4 relevant and readily understood with adequate support;3 mostly relevant and understandable but incompletely developed or noticeably inaccurate;2 poorly supported or partly relevant, with language that obscures ideas;1 few coherent ideas, severely restricted language or little original contribution;0 blank, non-English, rejected/unrelated topic, entirely copied prompt, or arbitrary keystrokes. Q109 task anchors: an opinion on breadth/specialization; explanations, examples or details supporting it; a contribution beyond merely repeating Mina or Daniel. A qualified combined approach, either student's position, or another supported position can succeed. Fewer than 100 words does not create an automatic zero. Listen and Repeat (Q110–116),each 0–5:5 fully intelligible exact repetition;4 meaning preserved despite minor word/grammar/pronunciation differences;3 essentially full sentence but meaning inaccurate, with most ideas retained and possible intelligibility trouble;2 large omissions or inaccuracies, fragmentary meaning or low intelligibility;1 minimal attempted repetition or largely unintelligible speech;0 no response, entirely unintelligible, no English, or unrelated content. Preserve negation, word order, conditions, referents, tense and number. A textual exact match cannot establish spoken 5. Meaning-preserving paraphrase is not exact repetition. Interview (Q117–120),each 0–5:5 fully addresses the question with developed content, clear fluent delivery, effective rhythm/intonation and precise varied language;4 addresses and develops it with generally clear delivery and adequate language;3 relevant but limited development/clarity, disrupted pace or restrictive language;2 minimally relevant, little elaboration and limited intelligibility/language;1 vague relevance, isolated language or mostly unintelligible speech;0 no response, entirely unintelligible, no English, or unrelated content. No preferred opinion is scored. Actual speech is required for oral judgments. Totals:Reading 50,Listening 47,Writing 20(10 objective+two 5 point tasks),Speaking 55(eleven 5 point tasks),maximum 172. Keep objective accuracy distinct from rubric ratings. None of these is an official raw total, scaled band, CEFR level, live-test pretest count, or pass prediction. Current sources conflict on some grand counts and raw R/L maxima; see official-source-conflicts.json and research.txt.