2027 G3 SEC English, Mathematics and Science | Learner’s Guide Vol 0180. A school exhibition announces, “Genetics guarantees three purple flowers for every four plants.” It sounds familiar because students have just learnt the 3:1 ratio from a simple monohybrid cross. But a prediction about probabilities is not a promise about an observed group. What if the school has counted forty plants and found twenty-seven purple? What if its classroom simulation always gives three purple in every four draws? Which result should readers trust?
This original EMS workshop follows the previous English, probability-tree and Biology chapters. It uses a fictional plant and a paper-card demonstration, not real school genetic data or an instruction to conduct unsupervised crosses. The purpose is to write a source-faithful report, calculate expected and observed ratios, test whether a simulation follows the biology model, and make a cautious recommendation that another student could check.
1. First, place the case in the real 2027 SEC curriculum
The official SEAB G3 2027 subject directory, last updated 26 October 2026, lists English Language K300, Mathematics K310, and Science combinations K326, K327 and K328. The English K300 syllabus assesses comprehension, clear written explanation and purpose-specific writing. The Mathematics K310 syllabus includes simple combined probability events and the use of possibility or tree diagrams when appropriate. The G3 Science syllabus covers monohybrid crosses, predicted ratios and explanations for observed differences within Biology.
Biology is part of Science K327 (Physics, Biology) and K328 (Chemistry, Biology), rather than K326 (Physics, Chemistry). In the shared Science assessment, Biology is Paper 4 for the appropriate subject pairings, lasting 1 hour 15 minutes and carrying 65 marks (32.5 per cent of the total). Paper 1 multiple choice is one hour, 40 marks (20 per cent), while Paper 5 practical is 1 hour 30 minutes, 30 marks (15 per cent). Candidates take the papers appropriate to their subject pairing.
There is no official combined English–Mathematics–Science exam paper called EMS. This fictional school exhibition is a transfer workshop: it helps students use literacy, calculation and scientific judgement on the same information. For the live exam, practise each subject’s actual format and instructions.
2. Read the fictional exhibition sources separately
Source A: Exhibition banner. “Our purple-petal gene test is perfectly accurate. Three quarters of the next generation will be purple. In every group of four seeds, three will show purple and one will show white. Our card game demonstrates the rule because it always produces a 3:1 result.”
Source B: Biology model note. “The exhibition uses an imaginary diploid plant and one gene with two alleles. T is completely dominant for purple petals; t is recessive for white petals. The modelled parent cross is Tt × Tt. Gametes from either parent have equal chances of carrying T or t. Different offspring are treated as independent outcomes for the probability exercises. Environmental factors, selection and other genetic effects are excluded unless stated.”
Source C: First observation table. “Forty offspring were classified in this fictional teaching record: 27 purple and 13 white. A separate log notes that another eight seeds in the same planting group did not produce a classifiable flower by the observation date. Their petal category was not recorded.”
Source D: Paper-card demonstration. “The students made four equally sized cards labelled TT, Tt, tT and tt. For each visitor, they shuffled the four cards, drew them one at a time without putting a card back, and labelled the first three purple and the last one white by their genotype. They displayed the four-card result as proof that every group of four offspring has a 3:1 phenotype ratio.”
Source E: Second model record. “A different fictional cross, Tt × tt, produced twenty-four classifiable offspring: 14 purple and 10 white. The exhibition plans to compare this data with the same 3:1 banner unless the model is changed.”
3. Read source type before calculating
The banner is a public claim. The model note supplies the genetic assumptions needed for calculation. The observation table reports a selected set of counted fictional plants. The card demonstration describes a procedure, which may or may not represent independent offspring. The second record belongs to a different parental cross and therefore requires a different expected ratio.
Do not combine these sources into one reassuring story before inspecting each. The fact that a statement appears on a school banner does not make it scientific evidence. Conversely, a table with a small deviation from expectation does not automatically disprove a probability model. Every source answers a different question.
A dependable source-to-answer method has three passes: identify what each source directly says, calculate what logically follows under the assumptions, and check whether the proposed public claim is broader than those results. Only then draft the report.
4. Translate the Biology model into possible outcomes
Under Tt × Tt, each parent can contribute T or t with probability one-half in the stated idealised model. Combining gametes gives four equally likely ordered outcomes: TT, Tt, tT and tt. Grouping the two mixed orders as the heterozygous genotype yields genotype probabilities 1/4 TT, 1/2 Tt and 1/4 tt.
Under complete dominance, TT and Tt show purple; tt shows white. Therefore the expected phenotype probabilities are 3/4 purple and 1/4 white. These are proportions in the model, not exact observed counts for every set of four seeds.
The Volume 0179 Biology guide explains the allele meanings and genetic diagrams in depth. The present case is about what happens when those predictions are reported publicly and compared with finite samples.
5. Why the 3:1 banner is too certain
The model predicts a purple phenotype probability of 3/4 for any one independent offspring under its assumptions. It does not stipulate that the first, second and third offspring must be purple or that every fourth must be white. Each new outcome is uncertain; a small batch can contain several different mixtures.
For exactly four independently modelled offspring, all four purple is possible with probability (3/4)⁴ = 81/256, approximately 31.6 per cent. A group of four containing no white is therefore not an impossible exception. The exact arrangement of three purple and one white is also only one of several possible categories of outcomes.
A more accurate banner would say, “Under a simple Tt × Tt complete-dominance model, each offspring has an expected 75% chance of showing purple petals; actual groups may differ.” This does not weaken the science. It protects the distinction between a theoretical chance and an observed count.
6. The probability-tree structure behind the phenotype result
Imagine a two-branch first stage representing a T allele from one parent or a t allele from that parent. Each branch has probability 1/2. From each, add a second-stage branch representing T or t from the other parent, also with probability 1/2 under the model. Each ordered combination has probability 1/4.
Multiplying along a path gives TT 1/4, Tt 1/4, tT 1/4 and tt 1/4. Adding the three paths that produce purple gives 3/4. This is the same branch-and-path discipline developed in Volume 0178.
The mathematics helps organise possibilities. It does not independently prove the genetic assumptions about equal gamete chances or complete dominance. Scientific evidence or the examination’s given statements supply those premises. Mathematics then shows what they imply.
7. Expected counts for the first observed group
Source C reports forty classified offspring. Under the Tt × Tt model, the expected purple count for forty independent outcomes is 40 × 3/4 = 30. The expected white count is 40 × 1/4 = 10. The expected ratio is 30:10, which simplifies to 3:1.
The observed table, however, contains 27 purple and 13 white. These figures still total forty. The observed phenotype ratio is 27:13, not 3:1. It should not be secretly rewritten as 30:10 merely because the latter looks like the textbook model.
The difference from expected counts is three fewer purple and three more white in the classified set. A sound report states both predicted and observed values, names their different sources, and avoids treating a small discrepancy as either perfect confirmation or decisive refutation.
8. Convert the observed counts to percentages carefully
The observed purple proportion is 27/40 = 67.5%. The observed white proportion is 13/40 = 32.5%. The model predictions are 75% and 25%, respectively. The observed purple percentage is 7.5 percentage points lower than predicted.
A percentage-point difference is not the same as a relative percentage change. Dividing 7.5 by 75 would give a 10 per cent relative shortfall from the predicted purple proportion, a different comparison. The most accessible report for school readers is often “67.5% observed, compared with 75% predicted”, followed by the 7.5-percentage-point difference.
Keep the denominator visible. It is forty classified offspring, not all forty-eight planted seeds mentioned in the separate log. Changing the denominator without adding valid classifications would misrepresent the data.
9. The eight unclassified seeds are not eight white flowers
Source C mentions eight seeds that did not yield a classifiable petal colour by the observation date. They are missing phenotype observations. A student might be tempted to include them as white because “no purple was seen”. That is invalid: lack of a recorded purple flower is not evidence of a white phenotype.
The classified purple proportion is 27/40 = 67.5%. A calculation using 27/48 = 56.25% could describe the fraction of all planted seeds known to have produced a purple-classified offspring, but it is not an observed purple phenotype proportion among forty-eight fully classified offspring. The eight missing classifications leave a gap.
This is a measurement-denominator lesson. What a data set can show depends on which items were observed, which were excluded and why. Before comparing a predicted genetic ratio with observations, record the status of unclassified or missing cases.
10. Simple logical bounds do not magically fill missing information
Suppose, only for an illustrative classification thought experiment, that each of the eight missing seeds could ultimately be assigned either the model’s purple category or white category. Among all forty-eight outcomes, the purple-category count could then range from the already observed 27 to at most 35. That gives a purely logical range from 27/48 = 56.25% to 35/48 ≈ 72.9%.
These are bounds on possible eventual category proportions under that hypothetical classification assumption. They are not the observed proportion now, nor a statistical confidence interval, nor an estimate that the unclassified seeds will all produce flowers. If some never develop classifiable petals, the thought experiment’s premise fails.
Why include the calculation? It shows that missing observations can affect a conclusion materially. Rather than hiding eight cases or inventing their phenotypes, a careful committee would report that classification is incomplete and explain the conditions under which any later totals are computed.
11. Why small groups depart from predicted ratios
Random sampling variation means a finite set of outcomes can have proportions different from the model’s long-run expectation. Under the independent offspring model, every new offspring has a 3/4 purple probability; this is compatible with stretches of all-purple or white-heavy results.
Forty observations are more informative than four in some respects, but even forty do not guarantee 30 purple and 10 white. Deviations can arise without a mistake in the Mendelian calculation. The official G3 Science Biology outcomes ask learners to explain why observed ratios may differ from expected ratios, especially for small numbers of offspring.
Do not claim that chance definitely caused the discrepancy. The packet does not establish whether all offspring were observed, whether classifications were reliable, or whether environmental conditions were uniform. Chance variation is one plausible explanation under the model; missing data and observational bias deserve separate consideration.
12. The four-card demonstration has a hidden design flaw
Source D uses four cards representing the four ordered allele combinations from Tt × Tt. The students then draw every card without replacement. Because the physical pack contains exactly three purple-labelled outcomes and one white-labelled outcome, every complete four-card demonstration will show three purple and one white.
That guarantee comes from the card-pack construction, not from independent biological offspring. It is like drawing all four different tickets from a four-ticket envelope: you will see every ticket once. The experiment is no longer a faithful model of four independent offspring whose genotype combinations can repeat.
The card game can demonstrate the four possible combinations and their equal probabilities for one draw. To simulate repeated independent offspring, replace the card after each draw, shuffle thoroughly and draw again under the stated model. The independent-draw assumption needs to be reflected in the procedure.
13. Two kinds of classroom activity should not be confused
Activity one is an enumeration exercise: lay out TT, Tt, tT and tt once each and count the three purple-labelled outcomes and one white-labelled outcome. This demonstrates how the probability of purple is built from possible combinations. It does not simulate a group of four independent offspring.
Activity two is a random sampling exercise: draw one card, record the outcome, replace it, mix and repeat as independent trials. Now groups of four can contain anywhere from zero to four purple results. Over many draws, the observed fraction can be compared with the model’s 3/4 prediction.
Both activities are legitimate when their purposes are named. The exhibit’s mistake is treating the exact composition of a complete no-replacement card pack as if it proved an exact four-offspring biological quota.
14. How likely is four purple in a row?
With replacement and independent draws from four equally likely cards containing three purple-labelled outcomes, the chance of purple on each draw is 3/4. The chance of four purple draws is (3/4)⁴ = 81/256 ≈ 31.6%.
This result is a good diagnostic. The flawed no-replacement demonstration makes all-purple impossible when every one of its four cards must be drawn. The corrected independent model gives all-purple a substantial possibility. If a simulation can never generate an outcome that the biological probability model permits, its mechanism is misaligned with the intended model.
The explanation matters more than the percentage’s last decimal. The experiment’s sampling rule—replacing or not replacing after each draw—changes the event structure. A correct report must describe that rule before presenting simulated frequencies.
15. What is the chance of at least one white in four draws?
The complement of “at least one white” is “all purple”. Since the all-purple probability is 81/256, P(at least one white) = 1 − 81/256 = 175/256, or about 68.4%.
This does not imply that a white result is guaranteed in a group of four. Nor does it contradict the 3:1 expectation. It is a statement about the chance that the group contains one or more white outcomes, regardless of exactly how many.
A probability-tree reader can explain it without a memorised formula: each purple-only path has a 3/4 chance at each stage under the independent model; multiplication gives the complement’s joint probability, and subtraction gives the requested at-least-one event.
16. A three-offspring example keeps path reasoning visible
For three independent offspring, P(all purple) = (3/4)³ = 27/64. Therefore P(at least one white) = 1 − 27/64 = 37/64, approximately 57.8%. Exactly two white requires three ordered arrangements: WWP, WPW and PWW.
Each of those paths has probability (1/4)(1/4)(3/4) = 3/64. Their sum is 9/64, about 14.1%. Students who learned to add successful complete paths in Volume 0178 will recognise why the three placements matter.
This is enrichment in straightforward combined-event probability reasoning, not a claim that the SEC Biology syllabus requires a named binomial distribution method. The underlying skill is to list valid outcomes, apply given chances and distinguish a prediction from a count.
17. Source E has a different parental cross
The second record belongs to Tt × tt, not Tt × Tt. Under the same model, the expected genotype combinations are Tt and tt in equal proportions; the phenotype expectation is 1 purple : 1 white. You cannot attach the 3:1 banner to this cross simply because it also concerns the fictional petal-colour gene.
With twenty-four independent modelled offspring, the expected counts are 12 purple and 12 white. The reported fictional observations are 14 purple and 10 white. That is 14/24 ≈ 58.3% purple, versus 50% expected. The difference from the model is about 8.3 percentage points.
This is a second lesson in denominators and comparison groups. The first table has forty classified observations and a 3:1 model. The second has twenty-four observations and a 1:1 model. Combining their counts before acknowledging the different crosses would create an overall percentage with no single simple expected ratio.
18. Why pooling the two records is misleading
Adding the raw counts gives 27 + 14 = 41 purple and 13 + 10 = 23 white, a total of sixty-four classified offspring across the two models. The combined observed purple proportion is 41/64 ≈ 64.1%. That number is arithmetically correct for the pooled observations but does not describe one uniform breeding scenario.
The expected combined purple count under the two different stated models would be 30 + 12 = 42, with expected white 10 + 12 = 22. The expected pooled fraction is 42/64 = 65.625%. But even this pooled expectation is a weighted combination of distinct crosses, not a new Mendelian phenotype rule.
A careful report shows both groups separately and explains why they differ in model. The aggregate can be included for an administrative question if needed, with an explicit warning that the source populations are not identical.
19. The same observed percentage can answer different questions
An observed 67.5% purple in the first group might be compared with 75% as a model check. It might also be compared with a school target such as “at least 60% purple” if a hypothetical display rule requested that. Those are distinct questions: one evaluates agreement with a genetic expectation, and the other checks a chosen operational threshold.
The same dataset can pass a threshold while deviating from a theoretical expectation. There is no contradiction. The comparison base and purpose differ. Students should label whether a conclusion concerns genetics, event feasibility, a target or a promotional claim.
This is the kind of language precision developed through earlier EMS evidence workshops. A calculation alone is not a conclusion until you state what decision or assertion it actually supports.
20. Revisit the exhibition banner line by line
Claim one, “Our gene test is perfectly accurate,” is unsupported. The fictional packet defines a model and reports samples, but it gives no independent validation showing perfection. Claim two, “Three quarters of the next generation will be purple,” is too categorical if it means exactly three-quarters of any finite offspring group. A probability prediction would be more appropriate.
Claim three, “Every group of four has exactly three purple and one white,” is false under the independent model. Claim four, “The card game proves the rule,” confuses a fixed four-card enumeration with sampling repeated independent outcomes. Correcting each statement is a substantive scientific task, not merely a punctuation exercise.
A responsible rewritten banner can be short: “In our simplified Tt × Tt model, each offspring has a 75% predicted chance of purple petals. Realised proportions can vary. Our labelled cards illustrate the four possible allele combinations; independent repeated draws require replacement between draws.”
21. Sentence boundaries make the report auditable
An awkward draft says, “The predicted ratio is three to one, the observed sample contains more white flowers, the rule is wrong.” This is a comma splice and an unsupported inference. A clearer version is: “The model predicts a 3:1 phenotype ratio. The classified sample contained 27 purple and 13 white offspring. The difference may reflect sample variation or other untested influences; it does not, by itself, disprove the model.”
The full stops separate model prediction and observation. The semicolon links a limited conclusion with the reason for caution. The modal may keeps untested explanations from becoming asserted causes. This is how grammatical accuracy supports scientific honesty.
For sentence-boundary repairs, consult Volume 0177. The scientific conclusion depends on the correct maths and Biology, but its readers depend on the prose.
22. Write a two-paragraph results note
Model note, paragraph one: “The exhibition uses a hypothetical Tt × Tt cross, which predicts purple in 75% and white in 25% of independent offspring. Among forty classifiable fictional offspring, twenty-seven were purple and thirteen white, or 67.5% and 32.5%. These are observed proportions, not the expected 3:1 values.”
Model note, paragraph two: “The four-card activity lists the possible gamete combinations, but drawing all four cards without replacement artificially guarantees three purple outcomes and one white. For an independent-outcome simulation, each card should be returned and the deck mixed before the next draw. Eight additional seeds were unclassified, so the sample record is incomplete. We recommend describing the model as a prediction rather than a guaranteed result.”
Both paragraphs are original examples. They distinguish observations, model calculations, a method flaw and a specific improvement. A complete situational writing task might require further points, recipient details and a suitable format; this compact report is training for evidence-based explanation, not a substitute for the K300 task.
23. How to improve the classroom simulation without chemicals
First decide the purpose. If the aim is only to teach the four possible offspring genotypes, lay out the four cards and count them. If the aim is to demonstrate variability across repeated independent offspring, draw one card at a time with replacement and mixing under an appropriate model. Record each result in a transparent tally.
Do not prearrange exactly three purple cards for each four-round visit and then present the tally as a genuine random sample. A demonstration that engineers the expected ratio every time hides the very variation the Biology syllabus asks students to understand.
Paper cards or a digital random model are sufficient for this classroom exercise. There is no need to handle real seeds, breed plants or conduct experiments outside supervised school settings. For an educational simulation, the central question is whether its probability mechanism matches the stated model.
24. A fair test needs a defined observation protocol
For a real biological observation—not performed in this fictional workshop—a school investigation would need to define the trait category, time of observation, treatment of unclassifiable individuals and conditions affecting survival. In an illustrative data audit, simply calling the count “forty offspring” is not as informative as stating “forty classifiable offspring, with eight additional unclassified seeds”.
Check whether observations were made before or after a selection step. If only germinated seedlings were counted, the inference concerns that observed subset. A difference between model prediction and observed ratio might warrant further data rather than instant rejection. Each interpretation should state its denominator and what the sample represents.
This is related to the Volume 0164 lesson on assignment versus delivery bias: people or items missing from observed results can change the meaning of an apparent percentage.
25. Data integrity before presentation design
A colourful exhibit can persuade an audience even when its denominator is wrong. The remedy is not a more attractive chart. The first remedy is to reconcile the source data: forty classified outcomes, eight unclassified entries, the parental cross in each group, and the process used in the simulation.
Then decide what each figure should show. A model panel might show the four equally likely genotype combinations for Tt × Tt; an observation panel might show the 27/13 classified sample. A note can explain that the two are not expected to match exactly. Separate panels for the second cross prevent the 1:1 expectation from being mistaken for 3:1.
In the public article, words and numbers are enough. Students who later build a diagram should ensure that colours, labels and legends do not conceal the distinction between predicted and observed data.
26. Thirty original integrated practice questions with worked answers
Read the fictional sources before each calculation. Unless a question changes the setup, use the Tt × Tt complete-dominance model with independent offspring, and treat unclassified outcomes as unknown rather than white. All data remain teaching illustrations.
01 · Subject code
Question. Which 2027 Science pairings contain Biology?
Worked answer. K327 and K328. Reason. K326 pairs Physics and Chemistry; Biology is not a component of that combination.
02 · Gametes
Question. Which alleles can a Tt parent supply?
Worked answer. T and t. Reason. Each with probability one-half in the simplified model.
03 · Genotypes
Question. List the four ordered Tt × Tt combinations.
Worked answer. TT, Tt, tT and tt. Reason. Each parent contributes one allele; the mixed pair occurs in two orders.
04 · Purple probability
Question. Under complete dominance, find P(purple).
Worked answer. 3/4. Reason. Three of the four equally likely ordered combinations express purple.
05 · White probability
Question. Find P(white).
Worked answer. 1/4. Reason. Only tt is white under the model.
06 · Predicted ratio
Question. What is the expected phenotype ratio from Tt × Tt?
Worked answer. 3 purple : 1 white. Reason. This ratio describes modelled probabilities, not fixed counts.
07 · Expected purple
Question. With 40 modelled offspring, what purple count is expected?
Worked answer. 30. Reason. 40 × 3/4.
08 · Expected white
Question. With 40 modelled offspring, what white count is expected?
Worked answer. 10. Reason. 40 × 1/4.
09 · Observed total
Question. How many classifiable offspring appear in Source C?
Worked answer. 40. Reason. 27 purple + 13 white.
10 · Observed purple
Question. What percentage of the classified group was purple?
Worked answer. 67.5%. Reason. 27/40 × 100.
11 · Observed white
Question. What percentage of the classified group was white?
Worked answer. 32.5%. Reason. 13/40 × 100.
12 · Percentage-point gap
Question. How far below the 75% prediction is the observed purple percentage?
Worked answer. 7.5 percentage points. Reason. 75 − 67.5.
13 · Missing cases
Question. Should the eight unclassified seeds be called white?
Worked answer. No. Reason. Missing petal classification is not evidence of a white phenotype.
14 · Known-purple among planted
Question. What percentage of all 48 planted seeds are known to be purple-classified?
Worked answer. 56.25%. Reason. 27/48 × 100; this has a different denominator and meaning from 27/40.
15 · Hypothetical upper bound
Question. If all eight unknowns could ultimately be classified purple, what is the maximum possible purple count among 48?
Worked answer. 35. Reason. Add 27 recorded purple and eight hypothetical purple unknowns.
16 · Hypothetical upper percentage
Question. Convert that hypothetical maximum to a percentage of 48.
Worked answer. About 72.9%. Reason. 35/48 × 100; not an observed percentage.
17 · Flawed simulation
Question. Why does drawing all four labelled cards without replacement yield exactly 3:1 every round?
Worked answer. The fixed deck contains exactly three purple-labelled and one white-labelled outcome. Reason. Sampling all four exhausts the predetermined composition.
18 · Better simulation
Question. How should cards be handled between draws to model independent offspring?
Worked answer. Replace the card and mix before the next draw. Reason. This restores the same outcome probabilities under the simple model.
19 · All-purple four
Question. Find P(all four purple) with independent draws.
Worked answer. 81/256. Reason. Multiply (3/4)⁴.
20 · At least one white
Question. Find P(at least one white in four independent outcomes).
Worked answer. 175/256. Reason. Take 1 minus 81/256.
21 · All-purple three
Question. Find P(three purple in three independent offspring).
Worked answer. 27/64. Reason. Multiply (3/4)³.
22 · At least one white of three
Question. Find P(at least one white among three).
Worked answer. 37/64. Reason. 1 − 27/64.
23 · Exactly two white of three
Question. Find P(exactly two white among three).
Worked answer. 9/64. Reason. There are three paths, each with chance (1/4)²(3/4) = 3/64.
24 · Second cross
Question. What is the phenotype prediction from Tt × tt?
Worked answer. 1 purple : 1 white. Reason. The expected probabilities are 1/2 each.
25 · Second expected purple
Question. What purple count is expected among 24 such offspring?
Worked answer. 12. Reason. 24 × 1/2.
26 · Second observed purple
Question. What is the observed purple percentage for 14 of 24?
Worked answer. About 58.3%. Reason. 14/24 × 100, rounded to three significant figures.
27 · Combine records
Question. How many classified purple offspring appear across the two different crosses?
Worked answer. 41. Reason. 27 + 14, but the crosses have different expectations.
28 · Pooled observed proportion
Question. What is the pooled purple fraction across all 64 classified offspring?
Worked answer. 41/64, about 64.1%. Reason. Arithmetic pooling is possible, but it must not be confused with a single Mendelian expectation.
29 · Correct the banner
Question. Is “exactly three purple among every four” a valid claim?
Worked answer. No. Reason. A 3/4 probability for each independent offspring does not enforce a fixed group composition.
30 · Recommendation
Question. What should the exhibition team do before claiming the card game proves its prediction?
Worked answer. Label enumeration and simulation separately, use replacement for independent draws, and report observations with their denominators. Reason. These changes directly address the demonstrated design and reporting flaws.
27. Change one assumption and rebuild the answer
Suppose the first model is changed to TT × tt. Then all modelled offspring are Tt and purple, so there is no 3:1 expected phenotype ratio. The four-card deck TT, Tt, tT, tt is no longer a valid representation of the offspring combinations for that cross. A student who memorised the deck’s colour count would continue producing a neat but false result.
Now restore Tt × Tt but change the task from “probability of purple for one offspring” to “proportion of purple among the forty classified plants”. The first answer is 3/4 from the model; the second is 27/40 from the observations. The underlying cross is unchanged, but the quantity requested is different.
Finally change the observation deadline so the eight unclassified cases remain missing. Any report that includes them as known white has invented data. A strong student notices which assumption changed and revises only the conclusion that depends on it.
28. A useful revision routine for the three subjects
For English, rewrite five promotional claims with exactly the certainty their sources allow. For Mathematics, compute the cross probabilities, complementary events and correct denominators without premature rounding. For Biology, identify the parental genotypes, trace gamete contributions and explain expected-versus-observed discrepancies without treating dominance as frequency.
Then bring all three together in a one-page report. Every quantitative sentence should name its base. Every genetic conclusion should name its cross. Every observation should identify the group actually counted. Use a clear subject and complete sentence rather than compressing unrelated claims into a comma splice.
The test of learning is whether you can explain why the unmodified four-card demonstration is unsuitable as an independent-offspring simulation. If you can identify that flaw, calculate its consequences and recommend a specific change, you have made a genuine transfer across EMS.
29. Questions from students and parents
Does three purple to one white mean every four seeds must follow that pattern? No. The ratio describes the expected phenotype distribution under a stipulated model; independent outcomes can vary.
Why is the no-replacement deck wrong? It forces every fixed group of four draws to contain exactly the three purple-labelled cards and one white-labelled card in the deck. Independent offspring can repeat the same genotype combination.
Does 27 purple out of 40 disprove complete dominance? Not by itself. Small-sample variation, classification limits and other causes require careful consideration. The observation is a count, not a proof of mechanism.
Should the eight unclassified seeds be ignored forever? No. Record their status and understand why their missing classifications matter. Do not invent a colour for them.
Does this article describe an actual Singapore school’s genetics experiment? No. The plants, observations, prices and classroom display are fictional educational examples.
30. Follow the series through the correct owner pages
The three immediately preceding foundation chapters are Volume 0177 on sentence boundaries, Volume 0178 on probability trees, and Volume 0179 on monohybrid inheritance. This integrated case tests whether the three skills survive a misleadingly confident school display.
The G3 SEC learner guide is the wider examination pathway. The PSLE Learning Guide supports earlier foundations. The Secondary English Hub, Mathematics Hub and Science Hub provide the subject-owned learning routes.
For examination requirements, use the official K300 English syllabus, K310 Mathematics syllabus and G3 Science K326/K327/K328 syllabus. This chapter’s school display is not an official integrated assessment; it is a way to practise the reasoning skills that those separate subjects cultivate.
31. The final recommendation and the lesson it carries
The school exhibition should retain the genetic-cross illustration but change its public wording. It should describe 3:1 as an expected phenotype ratio from a defined Tt × Tt model, not a guaranteed composition of each four-offspring group. It should report the actual fictional sample as 27 purple and 13 white among forty classified offspring, note eight unclassified seeds separately, and refrain from inventing a cause for the difference.
The card activity should distinguish laying out four possible ordered combinations from sampling independent offspring. The first is an enumeration of outcomes; the second requires appropriate repeated random sampling, such as replacement and mixing between draws. The second model record, Tt × tt, should have its own 1:1 expectation and not be folded into a universal 3:1 claim.
This is how careful learners progress: state the model, calculate the expectation, identify what was actually observed, check whether the method supports the comparison and write the conclusion without overselling it. The result is a report that stays useful even when a future reader asks an awkward but fair question.
