Statistical Tests in H2 Biology Paper 4: chi-squared, t-test, and How to Interpret Both
14 Apr 2026, 00:00 Z
Reviewed by
Ezekiel Tan·Academic Advisor (Biology)
Want small-group support? Browse our A-Level Biology Tuition hub. Not sure which level to start with? Visit Biology Tuition Singapore.
Looking for the full lab practical series? Visit the H2 Biology Practicals.
Practical course certificate note
For practical, lab, and experiment courses, Eclat Institute may issue an internal Certificate of Completion/Attendance based on participation and internal assessment.
- This is an internal centre-issued certificate, not an MOE/SEAB qualification or accreditation.
- Recognition (if any) is determined by the receiving school, institution, or employer.
- For SEAB private candidates taking science practical papers, SEAB states you should either have taken the subject before or complete a practical course before the practical exam date.
View our sample certificate template (Current sample layout (design may be refined over time))
Planning a revision session? Use our study places near me map to find libraries, community study rooms, and late-night spots.
> **Q:** What does this guide cover?\
> **A:** It explains which statistical tests appear in H2 Biology Paper 4, when to apply chi-squared versus a t-test, how to carry out each one with numbers, and how to write conclusions in the language examiners award marks for.
> **TL;DR**\
> The SEAB 9477 syllabus explicitly requires chi-squared tests for genetics ratios and ecological sampling data. The t-test is not named in the syllabus, but it appears regularly in school-based Paper 4 investigations and some data-handling questions where examiners provide two sets of means and ask for a statistical comparison. Know both. Apply chi-squared to frequency or count data. Apply the t-test when comparing two sample means from measurement data. State your null hypothesis before doing any calculation, interpret the p-value at the 5% level, and phrase your conclusion using "reject" or "fail to reject the null hypothesis."
For the full Paper 4 lab sequence, use the [H2 Biology practical guide for 9477 Paper 4](https://eclatinstitute.sg/blog/h2-biology-experiments/H2-Biology-Practical-2026-Lab-Mastery-Guide) alongside this page.
_Status:_ SEAB H2 Biology (9477, first exam 2026) syllabus last checked 2026-04-14 (PDF last modified 2025-11-28). Chi-squared is explicitly listed in the mathematical requirements. T-test appears in school practical contexts and data-handling questions but is not named in the published syllabus; it is included here as preparation for investigations where examiners supply a t-value or ask you to compare means.
---
## 1 Syllabus context: what the 9477 paper actually requires
The 9477 syllabus specifies that candidates should be able to "present data graphically, calculate rates, and apply chi-squared tests to results from genetics or ecological sampling." [1]
Chi-squared is the only named statistical test in the published requirements. However, Paper 4 investigations and data-handling questions have presented scenarios where students are expected to recognise a t-test result and interpret it correctly, particularly in ecology fieldwork comparisons and enzyme rate comparisons between treatment groups. School-based practicals and past question styles from analogous A-Level biology syllabuses treat the t-test as part of the expected quantitative toolkit.
This guide covers both. You will not be surprised by either in an exam or internal assessment.
---
## 2 Choosing the right test
| Test | Data type | Typical H2 Biology context |
| ---- | --------- | -------------------------- |
| Chi-squared ($\chi^2$) | Count or frequency data, comparing observed vs expected | Genetics ratios (9:3:3:1, 3:1), ecology species counts across two habitats |
| Independent samples t-test | Measurement data, comparing two group means | Mean leaf length in sun vs shade, mean enzyme rate at pH 6 vs pH 8, ecology fieldwork measurement comparisons |
| Paired t-test | Measurement data, same subject measured twice | Before-after enzyme inhibitor experiments (use this instead of independent t-test when the same sample is measured under both conditions) |
The decision rule: if your data are counts of outcomes in categories, use chi-squared. If your data are measured values and you want to compare the average of two groups, use the t-test.
---
## 3 Chi-squared test: worked example
### Context
A genetics experiment crosses two heterozygous parents (AaBb x AaBb) and gives an expected 9:3:3:1 phenotype ratio. You observe: 120 round yellow, 38 round green, 36 wrinkled yellow, 6 wrinkled green, from 200 offspring total.
### Step 1: State hypotheses
$H_0$: There is no significant difference between the observed and expected phenotype frequencies. The ratio conforms to a 9:3:3:1 Mendelian expectation.
$H_1$: There is a significant difference between observed and expected frequencies.
### Step 2: Calculate expected values
For a 9:3:3:1 ratio from 200 offspring: $E = \frac{total \times ratio\; share}{16}$
| Phenotype | Observed ($O$) | Expected ($E$) | $O - E$ | $(O - E)^2 / E$ |
| --------- | ------------- | -------------- | ------- | --------------- |
| Round yellow (9/16) | 120 | 112.5 | +7.5 | 0.50 |
| Round green (3/16) | 38 | 37.5 | +0.5 | 0.01 |
| Wrinkled yellow (3/16) | 36 | 37.5 | -1.5 | 0.06 |
| Wrinkled green (1/16) | 6 | 12.5 | -6.5 | 3.38 |
| **Total** | **200** | **200** | | **3.95** |
### Step 3: Calculate chi-squared
$$\chi^2 = \sum \frac{(O - E)^2}{E} = 3.95$$
### Step 4: Degrees of freedom
$$df = (\text{number of categories}) - 1 = 4 - 1 = 3$$
### Step 5: Compare to critical value
At $df = 3$ and $p = 0.05$, the critical value from standard chi-squared tables is 7.815.
Since $\chi^2 = 3.95 \lt 7.815$, the result is not significant at the 5% level.
### Step 6: Write the conclusion
> The calculated $\chi^2$ value (3.95) is less than the critical value (7.815) at 3 degrees of freedom ($p = 0.05$). We fail to reject the null hypothesis. There is no significant difference between the observed and expected phenotype frequencies. The data are consistent with a 9:3:3:1 Mendelian ratio.
---
## 4 Independent samples t-test: worked example
### Context
A student measures the maximum leaf width (mm) of ivy plants growing in two habitats: a shaded woodland understory and an open sunny grassland. She wants to know whether there is a significant difference in mean leaf width between the two habitats.
| Shaded leaves (mm) | Sunny leaves (mm) |
| ------------------ | ----------------- |
| 52 | 38 |
| 48 | 41 |
| 55 | 35 |
| 50 | 44 |
| 54 | 39 |
| 47 | 37 |
| 53 | 42 |
### Step 1: State hypotheses
$H_0$: There is no significant difference in mean leaf width between shaded and sunny habitats ($\mu_{shaded} = \mu_{sunny}$).
$H_1$: There is a significant difference in mean leaf width between shaded and sunny habitats ($\mu_{shaded} \neq \mu_{sunny}$). This is a two-tailed test because you have no prior reason to predict which direction the difference will go.
### Step 2: Calculate means and standard deviations
**Shaded group** ($n_1 = 7$):
$$\bar{x}_1 = \frac{52 + 48 + 55 + 50 + 54 + 47 + 53}{7} = \frac{359}{7} = 51.3 \text{ mm}$$
**Sunny group** ($n_2 = 7$):
$$\bar{x}_2 = \frac{38 + 41 + 35 + 44 + 39 + 37 + 42}{7} = \frac{276}{7} = 39.4 \text{ mm}$$
For the standard deviations, calculate the variance of each group and take the square root. Working shown below:
Shaded deviations from mean (51.3): $+0.7, -3.3, +3.7, -1.3, +2.7, -4.3, +1.7$
Squared deviations: $0.49, 10.89, 13.69, 1.69, 7.29, 18.49, 2.89$
Sum of squared deviations: $55.43$
$$s_1^2 = \frac{55.43}{7 - 1} = 9.24 \quad \Rightarrow \quad s_1 = 3.04 \text{ mm}$$
Sunny deviations from mean (39.4): $-1.4, +1.6, -4.4, +4.6, -0.4, -2.4, +2.6$
Squared deviations: $1.96, 2.56, 19.36, 21.16, 0.16, 5.76, 6.76$
Sum of squared deviations: $57.72$
$$s_2^2 = \frac{57.72}{7 - 1} = 9.62 \quad \Rightarrow \quad s_2 = 3.10 \text{ mm}$$
### Step 3: Calculate the pooled standard deviation
When group sizes are equal or similar and variances are similar, use the pooled standard deviation formula:
$$s_p = \sqrt{\frac{(n_1 - 1)s_1^2 + (n_2 - 1)s_2^2}{n_1 + n_2 - 2}} = \sqrt{\frac{(6)(9.24) + (6)(9.62)}{12}} = \sqrt{\frac{55.44 + 57.72}{12}} = \sqrt{9.43} = 3.07 \text{ mm}$$
### Step 4: Calculate t
$$t = \frac{\bar{x}_1 - \bar{x}_2}{s_p \sqrt{\dfrac{1}{n_1} + \dfrac{1}{n_2}}} = \frac{51.3 - 39.4}{3.07 \times \sqrt{\dfrac{1}{7} + \dfrac{1}{7}}} = \frac{11.9}{3.07 \times 0.535} = \frac{11.9}{1.64} = 7.26$$
### Step 5: Degrees of freedom
$$df = n_1 + n_2 - 2 = 7 + 7 - 2 = 12$$
### Step 6: Compare to critical value
For a two-tailed test at $p = 0.05$ with $df = 12$, the critical t-value from standard tables is 2.179.
Since $t = 7.26 \gt 2.179$, the result is significant.
### Step 7: Write the conclusion
> The calculated t-value (7.26) exceeds the critical value (2.179) for $df = 12$ at the 5% significance level (two-tailed). We reject the null hypothesis. There is a significant difference in mean leaf width between shaded and sunny habitats ($p \lt 0.05$). Shaded leaves have a significantly greater mean width (51.3 mm) than sunny leaves (39.4 mm), which is consistent with the expectation that plants in low-light environments develop larger leaf laminae to maximise light capture.
---
## 5 Interpreting the p-value
The p-value is the probability of observing a difference this large (or larger) by chance if the null hypothesis were true.
- $p \lt 0.05$: the result is statistically significant at the 5% level. Reject $H_0$.
- $p \geq 0.05$: the result is not statistically significant at the 5% level. Fail to reject $H_0$.
"Fail to reject" does not mean "accept" or "prove $H_0$." It means the data do not provide enough evidence to reject it.
The 5% level is the default in H2 Biology unless the question specifies otherwise (some questions specify 1% or 10%). Always state the level explicitly in your conclusion.
---
## 6 Writing conclusions: the examiner-ready phrasing
These sentence structures earn marks. Learn them exactly.
### For chi-squared (result is significant)
> The calculated $\chi^2$ value ([calculated value]) exceeds the critical value ([critical value]) for [df] degrees of freedom at the 5% significance level. We reject the null hypothesis. There is a significant difference between the observed and expected [frequencies/ratios]. The [describe the pattern] is not consistent with a [expected model].
### For chi-squared (result is not significant)
> The calculated $\chi^2$ value ([calculated value]) is less than the critical value ([critical value]) for [df] degrees of freedom at the 5% significance level. We fail to reject the null hypothesis. There is no significant difference between the observed and expected [frequencies/ratios]. The data are consistent with [expected model].
### For t-test (result is significant)
> The calculated t-value ([calculated value]) exceeds the critical value ([critical value]) for [df] degrees of freedom at the 5% significance level (two-tailed). We reject the null hypothesis ($p \lt 0.05$). There is a significant difference in mean [variable] between [group A] and [group B]. [Group A] has a significantly [greater/smaller] mean [variable] than [group B].
### For t-test (result is not significant)
> The calculated t-value ([calculated value]) is less than the critical value ([critical value]) for [df] degrees of freedom at the 5% significance level (two-tailed). We fail to reject the null hypothesis ($p \geq 0.05$). There is no significant difference in mean [variable] between [group A] and [group B].
---
## 7 Common mistakes
**Stating "the t-test is significant" without specifying the level.** A result is only significant at a stated confidence level. You must write "significant at the 5% level" or "significant at $p \lt 0.05$." Without the level, the conclusion is incomplete.
**Confusing one-tailed and two-tailed tests.** Use a two-tailed test when you have no prediction about the direction of the difference. Use a one-tailed test only when you have a clear biological reason to predict which group will have the higher mean, and you state this in $H_1$. In most Paper 4 scenarios, use the two-tailed critical value unless instructed otherwise.
**Not stating the null hypothesis before calculating.** Examiners award marks for the hypothesis statement independently of the calculation. Write $H_0$ and $H_1$ first, before any numbers.
**Writing "the hypothesis is proved."** Statistical tests never prove a hypothesis. They assess evidence. The conclusion is always about evidence supporting or not supporting $H_0$.
**Using "accept $H_0$" instead of "fail to reject $H_0$."** These are not equivalent. Failing to reject $H_0$ means the data are consistent with it. Accepting $H_0$ implies you have proved it true, which you cannot do from a finite sample.
**Applying t-test to categorical count data.** Chi-squared is for counts of things in categories (observed vs expected frequencies). T-test is for measured values where you are comparing means. Using t-test on count data gives meaningless results.
---
## 8 When not to use the independent samples t-test
**Use a paired t-test instead** when the same subjects are measured under two conditions (before and after a treatment, same plant measured in two seasons). The paired t-test accounts for within-subject variation and is more powerful in these designs.
**Use chi-squared instead** when your data are counts or frequencies of categorical outcomes, such as phenotype ratios or species presence/absence.
**Use Mann-Whitney instead** when you cannot assume the data are normally distributed (very small samples, obviously skewed distributions, or ordinal data such as rated scores). Mann-Whitney is the non-parametric equivalent of the independent t-test.
In H2 Biology, if a question provides a small dataset (fewer than 6 values per group) or clearly non-normal data, be cautious about applying t-test and note the assumption of normality as a limitation.
---
## 9 Summary: decision tree
1. Is your data in counts per category? Use **chi-squared**.
2. Are you comparing two means? Continue.
3. Are the two groups the same subjects measured twice? Use **paired t-test**.
4. Are the groups independent and data approximately normal? Use **independent t-test**.
5. Are the groups independent but data non-normal or very small? Use **Mann-Whitney**.
---
## Links and next steps
- To see chi-squared applied in the ecology fieldwork context with sampling tables: [H2 Biology Ecology Fieldwork Practical Guide](https://eclatinstitute.sg/blog/h2-biology-experiments/H2-Biology-Ecology-Fieldwork-Practical-Guide)
- For the full Paper 4 lab sequence and assessment structure: [H2 Biology Practical Guide for 9477 Paper 4](https://eclatinstitute.sg/blog/h2-biology-experiments/H2-Biology-Practical-2026-Lab-Mastery-Guide)
- For the H2 Biology practical hub: [H2 Biology practicals, labs, and experiments](https://eclatinstitute.sg/blog/h2-biology-experiments)
- For H2 Maths hypothesis testing context (z-test for population means, critical regions): [H2 Maths Notes: Hypothesis Testing](https://eclatinstitute.sg/blog/H2-Maths-Notes-6-5-Hypothesis-Testing)
---
> **Running a centre without lab facilities?**\
> We partner with private schools and homeschool centres to provide fully equipped labs, trained supervisors, and SEAB-aligned practical programmes. [Learn more](https://eclatinstitute.sg/blog/general/Science-Lab-Access-Private-Schools-Homeschool-Centres-Singapore)
## References
[1] SEAB. (2024). _Biology (Syllabus 9477) GCE A-Level 2026_ (first year of examination 2026, PDF last modified 2025-11-28). Singapore Examinations and Assessment Board. https://www.seab.gov.sg/files/A%20Level%20Syllabus%20Sch%20Cddts/2026/9477_y26_sy.pdf
[2] Campbell, N. A. et al. (2020). _Campbell Biology_ (12th ed.). Pearson - statistical analysis in biological investigations.
[3] Fowler, J., Cohen, L. and Jarvis, P. (1998). _Practical Statistics for Field Biology_ (2nd ed.). Wiley - t-test and chi-squared applications in ecological data analysis.
