# Educator-only full key and next teaching

These are authored desk models, not actual learner responses. Preserve first/later work separately. Equivalent defensible reasoning is welcome; a cautious uncertain prediction can be stronger than an unsupported ranking. Apply [task-specific criteria](RUBRIC.md), not a single score. Do not deliver this document to learners with their individual stage.

## Daily keys for all thirty route jobs

| Route | Accurate output or defensible reasoning | Targeted next teaching |
| --- | --- | --- |
| 1A | A assigned brighter/B shaded; temperature/moisture/actual light and adherence unknown; a chosen additional record tied to a specific question. | Give recorded/unrecorded headings and one condition at a time. |
| 1B | “Best” lacks criterion; every-bean future survival unanswerable; O2 recorded mean-increase question answerable. New question names living thing/condition/measure/period. | Repair one vague question using what/when/how measured. |
| 1C | Assignment recorded; universal growth unsupported; adherence unknown. An investigable question replaces overclaim. | Distinguish written intention from maintained control. |
| 2A | From Day 0: A +6/+12/+18 mm at Days 3/6/9; B +2/+4/+6 mm. Both 4 of 4 alive at Day 9 only. | Cover the final column and identify the initial 20 mm before subtraction. |
| 2B | x days 0/3/6/9, y height mm; all eight correct points, equal intervals, distinct labels/patterns. Lines do not reveal unobserved moments. | Plot one point with explicit pair/unit before connecting. |
| 2C | A increase 18 mm, not 38; shade killed B unsupported (both four alive); other two supported within fiction. | Separate growth, survival and proposed cause. |
| 3A | Actual repeated paper-line readings plus instrument/scale/reference, or supplied 28/43 mm labelled supplied; no prescribed real result. | Observe only the specific actual operation; correct zero placement, not invented data. |
| 3B | River soil-to-tip differs from Ridge pot-base-to-tip; shared method/unit/time/eye level recommended; outcomes unobserved. | Rename quantity rather than pool unlike readings. |
| 3C | 3 cm = 30 mm; 28 and 32 differ by 4 mm; coarse rounding can hide it; precision limited by method/instrument. | Compare ruler marks with reported number of digits. |
| 4A | Question varies assigned light while attempting equal species/stage/container/water/time; record height/alive/other conditions; shade-temperature confound remains. | Ask which feature is changed, which measured and which also might vary. |
| 4B | Defined day/time/sample/individual/reference rows; all future outcomes not observed; repetition permits pattern comparison without guarantee. | Add identifiers before summaries. |
| 4C | No unknown-plant collection/salt/public child images; supplied or adult-reviewed non-invasive plan; permissions cannot be presumed. | Convert command into request with explicit review role and no outcome. |
| 5A | Use E first/later full key below, capture first before outcome. | Choose just the failed component, then unused recheck after teaching. |
| 5B | 38−26 = 12 mm compares final means; each increase is 18 and 6 mm over nine days. | Build start/final/increase columns. |
| 5C | A relevant measured-light or temperature plan with specified instrument/unit/time, no actual reading; new data cannot retroactively repair old controls. | Separate prospective record from supplied historical fiction. |
| 6A | Actual contribution pair and a reason why complementary evidence/building on monitoring matters, from FACT1. | Ask what gap one contribution leaves; do not accept names alone as H01 explanation. |
| 6B | FACT3 historical development/difficulty/revised-test sequence and complementary expertise; anticipated harvest not a supplied completed result. | Label proposed future event separately from reported past work. |
| 6C | Camera alone insufficient; two source-backed contributions relevant to biological/measurement/design question and further evidence. | Connect expertise to an evidence need, not a generic praise of teamwork. |
| 7A | Tall criterion Q +16 vs P +11 mm; Alive criterion P 6/6 vs Q 4/6 at Day 8. Conditional, not a universal water recommendation. | State best for what before choosing. |
| 7B | Q 4/6, P 6/6; original six denominator retained; one check not lifetime. | Display all starting marks, including not-alive ones. |
| 7C | Relevant alternative condition question, consistent future labels/records; another explanation is unproved too. | Replace “caused” with source-bound observation plus testable question. |
| 8A | Real-source light-response link supports reviewed enquiry/possible coastal response, local unknowns and adult permission; no child electrical action. | Put place/period beside the scientific idea. |
| 8B | FACT2 historical not today/local/every reef; require named reef/date/method/response, no invented measures. | Circle current/local wording and request an actual source. |
| 8C | Place-specific monitoring better supports a reviewed response; turtle guidance from FACT4 not reef card. | Keep species, problem and source matched. |
| 9A | Own three-panel guide, source/sample/period, correct measures/representation/limit/next step; documented edit or justified keep. | Add just the missing reader-use component. |
| 9B | Criterion → relevant measure → limits → reviewed next enquiry; different criteria can yield P/Q; no automatic care claim. | Trace the same record with two explicit criteria. |
| 9C | Useful rewritten note with source-bound numbers and no universal/approved move; actual substantive edit recorded. | Replace one overclaim, then invite an own reason for the change. |
| 10A | Use F first/later full key below. New cohort remains separate from earlier cohort. | Reteach complete two-requirement table before choosing. |
| 10B | Actual project numbers/units/time/source audited and one unsupported claim improved, actionable next step. | Ask what someone could check next, not whether page looks attractive. |
| 10C | Specific question with measure/control/schedule/review, result that could challenge prediction hypothetical; outcome unobserved. | Write “would challenge” not “did happen”. |

## Ordinary model and home/supplement boundaries

C01–C20 are complete solved ordinary models in WORKED-CONTEXTS. H01–H10 and X01–X10 contain their complete supplied inputs, working/feedback and next job; they are open teaching material, not unused checks. The bridge own try is **11−8=3 mm over two days**, **2 of 3 alive at that check**. The local replacement model is mean 20 mm, increase 8 mm over five days, 3/4 alive. X10 is mean 19 mm, increase 8 mm over five days, 3/4 alive.

No ordinary/fresh model certifies actual plant responses, equipment operation, physiological health, cause, professional care, cultural permission or fieldwork. Where actual paper measurement occurs, accept the measured result and method/deviation; never mark it wrong for differing from invented lengths.

## Prediction-case common plan model

For E/RE first Q3: “How does the assigned condition relate to mean height change and alive-at-check count in these bean seedlings over a defined six-day period?” A paper plan identifies changed assigned condition, measures individual labelled soil-line-to-tip height in mm and separate alive count, keeps species/stage/container/other scheduled conditions/check time alike, checks Day 0/3/6, records relevant uncontrolled conditions and deviations. Risk/permission: educator reviews plant identity, safe non-invasive measurement, access and school/place permission before any actual observation; no collection or treatment is required. All outcomes in the plan are not observed. Do not claim that two controls automatically isolate causation.

For later Q4: unrecorded actual light/temperature/moisture or schedule adherence can weaken inference; ask for a consistent prospective record, same height reference/scale/time and relevant condition log. Supplied mean heights do not measure physiology or explain a death. The source's means cover the originally labelled sample, including not-alive stems with recorded heights; survivors-only summaries would be a different quantity and need individual identifiers.

## E · complete first answers

1. A `19−15=4 mm`, B `17−15=2 mm`, over three days. Assigned changed condition: **daily water amount**; actual controls/adherence remain unknown.
2. Source-supported possible prediction: **A** by the relative prior heights; justification uses the previous change. Accept a reasoned uncertain alternative explaining why earlier records do not guarantee Day 6. No future value or alive count exists first.
3. Use the common plan model with this case's condition; include all exact elicited components. Other relevant controls/questions accepted. A proposed plan is not actual I02 conduct/I03 measurement.

**Authored first model:** “The prior record gives A 4 mm and B 2 mm increase in three days. I predict A, if that relative pattern continues, but actual conditions are not recorded and growth may change. I would plan a consistent labelled height and alive-count record before making a care decision.”

## E · complete later answers

1. A `22−15=7 mm`; B `18−15=3 mm`, over six days. Day 6 relative mean: **A greater** (22/18 mm). Do not call final heights the increases.
2. An A-greater prediction can validly be kept with a source-bound limit; a qualified uncertainty prediction can now acknowledge A's listed mean is greater. No forced edit. Actual keep/edit must be shown in the learner's own later response, not inferred from this model.
3. A **4/4 alive at Day 6**; B **3/4**. These counts and mean-height changes do not establish health, later survival or a universal condition rule; original mean sample remains four.
4. Use a specific common limit/next question/improved rule. All record values are supplied, not enacted observation.

**Next teaching:** If future outcome is called first-known, practise a separate ordinary before/outcome envelope; if growth/survival blend, separate their columns. Then reserve RE1 or RE2 can be selected without prior preview.

## RE1 · complete first answers

1. A `14−10=4 mm`, B `15−10=5 mm`, over three days. Assigned changed condition: **assigned light position**; actual controls/adherence remain unknown.
2. Source-supported possible prediction: **B** by the relative prior heights; justification uses the previous change. Accept a reasoned uncertain alternative explaining why earlier records do not guarantee Day 6. No future value or alive count exists first.
3. Use the common plan model with this case's condition; include all exact elicited components. Other relevant controls/questions accepted. A proposed plan is not actual I02 conduct/I03 measurement.

**Authored first model:** “The prior record gives A 4 mm and B 5 mm increase in three days. I predict B, if that relative pattern continues, but actual conditions are not recorded and growth may change. I would plan a consistent labelled height and alive-count record before making a care decision.”

## RE1 · complete later answers

1. A `16−10=6 mm`; B `14−10=4 mm`, over six days. Day 6 relative mean: **A greater** (16/14 mm). Do not call final heights the increases.
2. A B-greater prediction based on Day 3 needs revision for the listed Day 6 result: A is now greater. B's mean decreases 1 mm from Day 3 to Day 6; possible error or biological change is unrecorded, so do not invent the cause. A prediction of uncertainty can update without claiming certainty. Actual keep/edit must be shown in the learner's own later response, not inferred from this model.
3. A **4/4 alive at Day 6**; B **4/4**. These counts and mean-height changes do not establish health, later survival or a universal condition rule; original mean sample remains four.
4. Use a specific common limit/next question/improved rule. All record values are supplied, not enacted observation.

**Next teaching:** If first prediction is retroactively erased, retain both versions and point to the new value that changed the comparison. If the mean drop is given an invented cause, make a possible-error question.

## RE2 · complete first answers

1. A `16−12=4 mm`, B `16−12=4 mm`, over three days. Assigned changed condition: **assigned potting-mix type**; actual controls/adherence remain unknown.
2. Source-supported possible prediction: **equal/uncertain** by the relative prior heights; justification uses the previous change. Accept a reasoned uncertain alternative explaining why earlier records do not guarantee Day 6. No future value or alive count exists first.
3. Use the common plan model with this case's condition; include all exact elicited components. Other relevant controls/questions accepted. A proposed plan is not actual I02 conduct/I03 measurement.

**Authored first model:** “The prior record gives A 4 mm and B 4 mm increase in three days. I predict equal/uncertain, if that relative pattern continues, but actual conditions are not recorded and growth may change. I would plan a consistent labelled height and alive-count record before making a care decision.”

## RE2 · complete later answers

1. A `19−12=7 mm`; B `19−12=7 mm`, over six days. Day 6 relative mean: **equal** (19/19 mm). Do not call final heights the increases.
2. An equal/tied or qualified uncertain prediction can be kept for mean height: both listed means are 19 mm. Their different alive counts still need separate reporting; a tie in height is not equality in survival. Actual keep/edit must be shown in the learner's own later response, not inferred from this model.
3. A **3/4 alive at Day 6**; B **4/4**. These counts and mean-height changes do not establish health, later survival or a universal condition rule; original mean sample remains four.
4. Use a specific common limit/next question/improved rule. All record values are supplied, not enacted observation.

**Next teaching:** If a tie forces an invented winner, contrast equal height with differing alive counts and state the criterion before deciding.

## F · complete first answers

1. A `32−20=12 mm`, **4/6 alive at Day 8**; B `29−20=9 mm`, **6/6 alive**; both over eight days, means over all six recorded labelled heights.
2. Both requirements apply: increase ≥10 mm **AND** alive ≥5 of 6.

| First option | Increase requirement | Alive requirement | Meets both |
| --- | --- | --- | --- |
| A | 12 mm: pass | 4/6: fail | fail |
| B | 9 mm: fail | 6/6: pass | fail |

Decision: **Neither** under the first reader's growth priority where more than one meets minima. A fails alive count; B fails height increase. Never substitute A growth/B survival as a third option.
3. Growth measure and alive count address different aims; the full rule uses both. Actual conditions/adherence unrecorded, small eight-day constructed cohort, not cause or future care. Ask for a reviewed comparable next cohort/condition log and labelled measures. This is a fictional proposed decision, not authentic community consultation or impact.

**Authored first note:** “For the stated two requirements, neither is my conditional decision, using 12/9 mm increases and 4/6 of six alive. We still lack actual condition logs; ask for a reviewed comparison before any real change.”

## F · complete later answers

1. Different new cohort: A `31−20=11 mm`, **6/6 alive**; B `34−20=14 mm`, **5/6 alive**, each at its own Day 8. This is not Day 16 of the old plants.
2. Same minima, later priority applied.

| New option | Increase requirement | Alive requirement | Meets both |
| --- | --- | --- | --- |
| A | 11 mm: pass | 6/6: pass | pass |
| B | 14 mm: pass | 5/6: pass | pass |

Later conditional decision **A**. Both now meet minima; A's six alive beats B's five under later survival priority, even though B has greater height increase.
3. Expected source-bound change **Neither → A**, linked to these new values/reader rule. If first answer differed, learner compares their actual captured choice and corrects or justifies it; do not fabricate the model's initial answer as theirs. New cohort starts again at 20 mm and has its own eight-day period.
4. A useful later note states A, includes both measures, acknowledges unknown actual conditions/adherence/future survival and proposes a reviewed evidence request. No universal treatment/care claim or actual permission/impact.

**Authored later model:** “For this different cohort and the later reader's complete rule, I recommend A conditionally. A's 11 mm increase and 6 of six alive support it. This changes my first decision because the supplied cohort and priority differ. Ask for consistent condition and individual records before an actual decision.”

**Next teaching:** Build a separate pass/fail cell for each condition; practise an ordinary neither-meets-both example. Then use a new ordinary both-eligible/different-priority case; only later select an unused RF case.

## RF1 · complete first answers

1. A `20−12=8 mm`, **6/6 alive at Day 8**; B `25−12=13 mm`, **5/6 alive**; both over eight days, means over all six recorded labelled heights.
2. Both requirements apply: increase ≥8 mm **AND** alive ≥5 of 6.

| First option | Increase requirement | Alive requirement | Meets both |
| --- | --- | --- | --- |
| A | 8 mm: pass | 6/6: pass | pass |
| B | 13 mm: pass | 5/6: pass | pass |

Decision: **B** under the first reader's growth priority where more than one meets minima. Both meet both minima, so B's greater increase supports the specified priority.
3. Growth measure and alive count address different aims; the full rule uses both. Actual conditions/adherence unrecorded, small eight-day constructed cohort, not cause or future care. Ask for a reviewed comparable next cohort/condition log and labelled measures. This is a fictional proposed decision, not authentic community consultation or impact.

**Authored first note:** “Both options meet the requirements. I choose B under the growth priority: A increased 8 mm with 6 of 6 alive; B increased 13 mm with 5 of 6 alive. We still lack actual condition logs; ask for a reviewed comparison before any real change.”

## RF1 · complete later answers

1. Different new cohort: A `22−12=10 mm`, **5/6 alive**; B `23−12=11 mm`, **4/6 alive**, each at its own Day 8. This is not Day 16 of the old plants.
2. Same minima, later priority applied.

| New option | Increase requirement | Alive requirement | Meets both |
| --- | --- | --- | --- |
| A | 10 mm: pass | 5/6: pass | pass |
| B | 11 mm: pass | 4/6: fail | fail |

Later conditional decision **A**. A meets both; B fails alive minimum despite greater height increase. Priority cannot rescue a failed minimum.
3. Expected source-bound change **B → A**, linked to these new values/reader rule. If first answer differed, learner compares their actual captured choice and corrects or justifies it; do not fabricate the model's initial answer as theirs. New cohort starts again at 12 mm and has its own eight-day period.
4. A useful later note states A, includes both measures, acknowledges unknown actual conditions/adherence/future survival and proposes a reviewed evidence request. No universal treatment/care claim or actual permission/impact.

**Authored later model:** “For this different cohort and the later reader's complete rule, I recommend A conditionally. A's 10 mm increase and 5 of six alive support it. This changes my first decision because the supplied cohort differs. Ask for consistent condition and individual records before an actual decision.”

**Next teaching:** A higher growth figure does not override a failed survival minimum. Keep eligibility separate from priority.

## RF2 · complete first answers

1. A `29−18=11 mm`, **4/6 alive at Day 8**; B `28−18=10 mm`, **5/6 alive**; both over eight days, means over all six recorded labelled heights.
2. Both requirements apply: increase ≥10 mm **AND** alive ≥4 of 6.

| First option | Increase requirement | Alive requirement | Meets both |
| --- | --- | --- | --- |
| A | 11 mm: pass | 4/6: pass | pass |
| B | 10 mm: pass | 5/6: pass | pass |

Decision: **A** under the first reader's growth priority where more than one meets minima. Both meet both minima; A's 11 mm increase is greater than B's 10 mm.
3. Growth measure and alive count address different aims; the full rule uses both. Actual conditions/adherence unrecorded, small eight-day constructed cohort, not cause or future care. Ask for a reviewed comparable next cohort/condition log and labelled measures. This is a fictional proposed decision, not authentic community consultation or impact.

**Authored first note:** “Both options meet the requirements. I choose A under the growth priority: A increased 11 mm with 4 of 6 alive; B increased 10 mm with 5 of 6 alive. We still lack actual condition logs; ask for a reviewed comparison before any real change.”

## RF2 · complete later answers

1. Different new cohort: A `28−18=10 mm`, **4/6 alive**; B `29−18=11 mm`, **5/6 alive**, each at its own Day 8. This is not Day 16 of the old plants.
2. Same minima, later priority applied.

| New option | Increase requirement | Alive requirement | Meets both |
| --- | --- | --- | --- |
| A | 10 mm: pass | 4/6: pass | pass |
| B | 11 mm: pass | 5/6: pass | pass |

Later conditional decision **B**. Both meet minima; B's five alive beats A's four under later survival priority.
3. Expected source-bound change **A → B**, linked to these new values/reader rule. If first answer differed, learner compares their actual captured choice and corrects or justifies it; do not fabricate the model's initial answer as theirs. New cohort starts again at 18 mm and has its own eight-day period.
4. A useful later note states B, includes both measures, acknowledges unknown actual conditions/adherence/future survival and proposes a reviewed evidence request. No universal treatment/care claim or actual permission/impact.

**Authored later model:** “For this different cohort and the later reader's complete rule, I recommend B conditionally. B's 11 mm increase and 5 of six alive support it. This changes my first decision because the supplied cohort and priority differ. Ask for consistent condition and individual records before an actual decision.”

**Next teaching:** Read the later reader's priority before ranking eligible options; a new priority is not a retroactive claim that first reasoning was wrong.

## Bounded next-teaching menu

- **Start/final confusion:** invented stem 7→11 mm over two days; model 4 mm increase, then learner tries 9→14 over three days (5 mm).
- **Denominator confusion:** show five original marks with three alive at one check; ask 3/5, not 3/3; then seven with four alive.
- **Controls versus claims:** a plan schedules equal water; a blank delivery log leaves adherence unknown. Ask for actual record, not an invented control.
- **Prediction versus hindsight:** ordinary 10→12 mm then unseen third check; collect a prediction first, reveal an authored different result separately, preserve both.
- **Criterion omission:** ordinary +6 mm/4 of 4 versus +9 mm/2 of 4 with minimum +5 and at least three alive: A only. Ask separate pass/fail before preference.
- **Mean sample confusion:** keep every original labelled height or explicitly name recorded/survivor subset; missing is not zero. Redefining quantity requires the task/key to change together.
- **Source versus local/live:** put provider, event, update/check and place next to a claim; if absent, write absent. Ask what primary body was actually read.
- **H01 names without reasoning:** join actual contribution to evidence need and explain why another contribution helps. Invented classroom roles do not demonstrate real scientific advances.
- **H02 action without evidence:** identify source-backed problem/science idea/response and local unknown before an adult-reviewed proposal. A task answer is not an enacted community decision.

After targeted ordinary teaching, select **one genuinely unused reserved case** whose job matches the remaining need. Do not use all four as repetitive retests or declare mastery from one later success. Actual learner evidence, physical conduct, real equipment and local consultation remain separate.