← A-Level Biology · CCEA · A-Level Biology revision
⭐ 0
CCEA GCE Biology (1010) · Section 4.7 Mathematical skills (statistics; taught content)
Mini-Lesson

Statistics and maths skills

5 missions. Guess, learn, then check. Your place is saved.

  1. Mission 1Mendel's peas came out at almost three to one, never exactly. How close is close enough?
  2. Mission 2Two sets of results share the same average but look nothing alike. What number captures the difference?
  3. Mission 3A difference between two groups could be real or just luck. How do biologists decide which?
  4. Mission 4Leaves in the shade look bigger than leaves in the sun. How big must the gap be before you believe it?
  5. Mission 5Two bars on a graph have different heights. When does that difference actually mean something?
Mission 1

Mission 1: Mendel's peas came out at almost three to one, never exactly. How close is close enough?

Make a guess: Mendel's peas came out at almost three to one, never exactly. How close is close enough?

Mission 1

Chance or a real difference?

  • Which gametes fuse is random. So observed counts scatter around the expected ones.
  • The χ² test asks: is the difference small enough to be due to chance?
  • Null hypothesis: there is no significant difference between observed and expected results.
  • The test uses counts in categories (phenotype classes). Never percentages or ratios.
  • The counted offspring are a sample of all the offspring the cross could produce.
Mission 1

The formula

  • χ² = Σ (O − E)² ÷ E. O = observed, E = expected, Σ = sum over all classes.
  • E = total counted × expected fraction for that class.
  • Degrees of freedom = number of classes − 1.
  • Big differences → big χ². A perfect fit → χ² = 0.
Mission 1

Worked example: Mendel's seed shape

Observed vs expectedMendel’s F2 seed shape0100020003000400050006000number of seeds54745493round18501831wrinkledO observedE expectedχ² = 0.26Mendel, 1866

F2 of round × wrinkled peas (Mendel, 1866). Null hypothesis: the data fit 3 round : 1 wrinkled.

RoundWrinkledTotal
O547418507324
E549318317324
(O − E)²/E0.0660.197
  • E (round) = 7324 × ¾ = 5493. E (wrinkled) = 7324 × ¼ = 1831.
  • χ² = 0.066 + 0.197 = 0.26.
  • Degrees of freedom = 2 − 1 = 1.
Mission 1

Critical value and decision

Decision line012345678910χ²critical value, p = 0.053.84 (1 df)significant:reject the null hypothesisnot significantaccept the null hypothesisMendel: χ² = 0.26
Degrees of freedom1234
p = 0.053.845.997.819.49
  • Find the critical value for your degrees of freedom at p = 0.05.
  • χ² below it: the difference is not significant. A difference this large would arise by chance more than 5% of the time.
  • So accept the null hypothesis: the difference is due to chance.
  • χ² above it: a significant difference. Reject the null hypothesis.
  • Mendel: 0.26 < 3.84. His data fit 3 : 1.
  • Say: "the data are consistent with 3 : 1" — not: "the data prove 3 : 1".
Quick check
?A student carries out a chi-squared test on the F2 offspring of a dihybrid cross. Which statement about the test is correct?
Mission 1Read the data

Mendel's dihybrid F2

Round yellow × wrinkled green; F1 self-pollinated (Mendel, 1866). Null hypothesis: 9 : 3 : 3 : 1.

PhenotypeOE(O − E)²/E
Round yellow315312.750.016
Round green108104.250.135
Wrinkled yellow101104.250.101
Wrinkled green3234.750.218
Total556556χ² = 0.47
  • E = 556 × 9/16, 3/16, 3/16, 1/16.
  • 4 classes → 3 degrees of freedom → critical value 7.81.
  • 0.47 < 7.81: not significant. The data fit independent assortment.
Quick check
?A cross between two double heterozygotes gives 160 F2 offspring: 95 show both dominant traits, 28 show the first dominant and the second recessive, 30 show the first recessive and the second dominant, and 7 show both recessive traits (illustrative data). The critical value of chi-squared at p = 0.05 is 7.815 for 3 degrees of freedom and 9.488 for 4 degrees of freedom. What is the outcome of a chi-squared test against a 9:3:3:1 ratio?
Mission 1

When the null hypothesis is rejected

  • A significant difference means the expected ratio is probably wrong for this cross.
  • Then look for a biological reason:
    • genes interacting (epistasis);
    • a lethal genotype;
    • a sex-linked gene;
    • unequal survival of one phenotype.
  • χ² tells you whether the difference is significant, not why.
Mission 1

Common slips

  • Use raw counts, not percentages.
  • Scale E to your total. Never use 9, 3, 3, 1 as E.
  • Degrees of freedom = classes − 1.
  • Big χ² = big difference. Above the critical value → significant.
  • Not significant ≠ proved.
Mission check

?In tomatoes, purple stems (A) are dominant to green stems (a). An Aa plant is crossed with an aa plant. Of 200 seedlings, 115 are purple and 85 are green (illustrative data). Critical values at p = 0.05: 3.84 for 1 degree of freedom, 5.99 for 2. Which conclusion is correct?
Mission 2

Mission 2: Two sets of results share the same average but look nothing alike. What number captures the difference?

Make a guess: Two sets of results share the same average but look nothing alike. What number captures the difference?

Mission 2

Three averages and a range

  • Mean = sum of values ÷ number of values.
  • Median = the middle value when the values are in order (for an even number, halfway between the middle two).
  • Mode = the most frequent value.
  • Range = largest − smallest.
  • One extreme value pulls the mean towards it. The median hardly moves.
Quick check
?The heights in centimetres of nine students are: 150, 152, 152, 155, 158, 160, 163, 170 and 175. What is the median height?
Mission 2

Standard deviation: the typical spread

Same mean, different SDShore ASD 1.2Shore BSD 6.5mean = 8mean ± 1 SD ≈ 68%± 2 SD ≈ 95%05101520Limpets per quadrat
  • Standard deviation (SD): how far values typically lie from the mean.
  • Small SD: values cluster tightly. Large SD: values are widely spread.
  • Two samples can share a mean and differ hugely in SD: the hook.
  • For data with a normal (bell-shaped) distribution, about 68% of values lie within ±1 SD of the mean, and about 95% within ±2 SD.
  • SD has the same unit as the data.
  • A large SD is not a "mistake". In biology it often reflects real variation between individuals.
Mission 2

Worked example: calculating SD

Stomata counted in five fields of view on one leaf: 22, 26, 19, 27, 21.

Count, xx − x̄(x − x̄)²
22−11
2639
19−416
27416
21−24
  • Mean, x̄ = 115 ÷ 5 = 23.
  • Σ(x − x̄)² = 46.
  • SD, s = √[Σ(x − x̄)² ÷ (n − 1)] = √(46 ÷ 4) = √11.5 = 3.4.
  • Divide by n − 1 (here 4), not n: this estimates the SD of the whole population from a sample.
  • CCEA writes this estimate as σ̂.
Quick check
?Limpets were counted in twenty random quadrats on each of two shores. On both shores the mean was 8 limpets per quadrat. The standard deviation was 1.2 on shore A and 6.5 on shore B. What can be concluded?
Mission 2

SD or range?

  • The range uses only two values: the largest and the smallest.
  • One outlying result sets the range on its own.
  • The SD uses every value, so it describes the spread of the whole data set.
  • So SD is usually the more useful measure of dispersion, especially where there is an outlying result.
  • Still check an outlier: is it an anomaly (a mistake) or real variation?
  • Say: "SD shows the spread of the data about the mean" — not: "SD shows how many errors were made".
Quick check
?Two samples of seven leaf lengths, in mm:
Both have a mean of 20 mm and a range of 20 mm. The standard deviation of P is 5.8 mm and of Q is 10.0 mm. Which statement is correct?
  • P: 10, 20, 20, 20, 20, 20, 30
  • Q: 10, 10, 10, 20, 30, 30, 30
Mission check

?The lengths of five seedling roots, in mm, are 14, 16, 15, 18 and 17. Which gives the mean and standard deviation correctly?
Mission 3

Mission 3: A difference between two groups could be real or just luck. How do biologists decide which?

Make a guess: A difference between two groups could be real or just luck. How do biologists decide which?

Mission 3

Could it just be chance?

  • Samples vary, so two groups almost never match exactly.
  • Start with the null hypothesis: there is no difference (or no association, or no correlation).
  • A statistical test turns the data into one number, the test statistic.
  • Compare it with the critical value at p = 0.05.
  • Above the critical value: if the null hypothesis were true, a result this extreme would happen by chance less than 5% of the time. The result is significant: reject the null hypothesis.
  • Below it: not significant. The difference could be due to chance.
  • Then accept the null hypothesis.
  • Significant does not mean proved; not significant does not prove "no effect".
Mission 3

Which test?

Which test?What kind of data?Counts incategoriesMeasurementsCompare withexpectedχ2Difference betweentwo means?Student's t-testRelationship betweentwo variables?scatterdiagram
  • Chi-squared (χ²): counts in categories, compared with expected counts.
  • Student's t-test: two means of continuous data (e.g. lengths, masses).
  • Ask: Counts or measurements? Difference or relationship?
  • t-tests and correlation coefficients are worked through in Mission 4.
Quick check
?A student measures the lengths of 20 leaves from plants growing in the sun and 20 from plants growing in the shade. She wants to know whether mean leaf length differs between the two. Which statistical test should she use?
Mission 3

Worked example: chi-squared beyond genetics

Daisies counted in two areas of a field of equal size, one trampled and one not (illustrative data). Null hypothesis: no difference, so equal numbers are expected.

TrampledUntrampledTotal
Observed, O3862100
Expected, E5050100
(O − E)² ÷ E2.882.88
  • χ² = Σ (O − E)² ÷ E = 2.88 + 2.88 = 5.76.
  • Degrees of freedom = categories − 1 = 1. Critical value at p = 0.05: 3.84.
  • 5.76 > 3.84: significant. Reject the null hypothesis: daisies are not spread evenly between the two areas.
  • The test does not say why. Trampling is one explanation to test further.
  • Expected values come from the null hypothesis: here, equal areas → equal numbers.
Quick check
?Fifty individuals of a hypothetical species of small ground-dwelling invertebrate were placed in the centre of a choice chamber with a moist side and a dry side. After 10 minutes, the number on each side was counted, as shown in the table. The null hypothesis is that the animals show no preference, so equal numbers are expected on each side. Using a chi-square test with p = 0.05, which conclusion is correct?
SideMoistDry
Number of animals3020
Mission 3

Correlation

Scatter graphsxystrong positivexyweak positivexyno correlationr = 0xystrong negative
  • Plot two measured variables on a scatter graph (scattergram).
  • Positive correlation: both rise together. Negative: one rises as the other falls.
  • A correlation coefficient runs from −1 to +1.
  • The sign gives the direction; the size gives the strength. 0 means no correlation.
  • −0.85 is as strong as +0.85.
Mission 3

Correlation is not cause

  • A correlation, even a strong one, can arise from:
    • chance (small samples);
    • a third variable linked to both (a confounding variable);
    • cause running the other way.
  • Evidence for cause: a mechanism; controlled experiments or randomised controlled trials; the effect grows with the dose; many studies agree.
  • Smoking and lung cancer: a strong correlation, plus a mechanism (carcinogens in tar damage DNA) and animal experiments. That combination is why it is accepted as causal.
  • Say: "there is a positive correlation between X and Y" — not: "X causes Y", unless there is evidence for cause.
Quick check
?A 1999 study found that children who had slept with a light on before the age of two were more likely to become short-sighted later. Later studies found that short-sighted parents were more likely to leave a light on at night, and short sight is partly inherited. Which conclusion is best supported?
Mission check

?A birdwatcher records 90 visits by finches to three feeders hung at different heights. Null hypothesis: the finches show no preference, so equal numbers of visits are expected.
FeederLowMiddleHigh
Visits423018

Critical values of χ² at p = 0.05: 1 degree of freedom, 3.84; 2, 5.99; 3, 7.81. Which conclusion is correct?

Mission 4

Mission 4: Leaves in the shade look bigger than leaves in the sun. How big must the gap be before you believe it?

Make a guess: Leaves in the shade look bigger than leaves in the sun. How big must the gap be before you believe it?

Mission 4

Student's t-test

What makes t bigx̄1x̄2small tx̄1x̄2bigger tx̄1x̄2bigger stillsample 1sample 2t =|x̄1 − x̄2|√(σ̂12/n1 + σ̂22/n2)gap between meansspread ÷ sample size
  • Use it to compare two means of continuous data that are roughly normally distributed.
  • Null hypothesis: there is no significant difference between the two means.
  • t gets bigger when the means are further apart, the SDs are smaller, or the samples are larger.
  • t = |x̄1 − x̄2| ÷ √(s1²/n1 + s2²/n2)
  • Degrees of freedom = n1 + n2 − 2.
  • t ≥ critical value at p = 0.05: reject the null hypothesis. The difference is significant.
  • The statistics sheet writes this with σ̂ (the estimated SD), and gives the t table.
Mission 4

Worked example: t-test

Limpet shell height on an exposed and a sheltered shore (illustrative data), 10 limpets each.

ShoreMean height, x̄ / mmSD, s / mmn
Exposed18.22.110
Sheltered15.62.410
  • Difference in means = 18.2 − 15.6 = 2.6 mm.
  • s1²/n1 = 2.1² ÷ 10 = 0.441. s2²/n2 = 2.4² ÷ 10 = 0.576.
  • √(0.441 + 0.576) = √1.017 = 1.008.
  • t = 2.6 ÷ 1.008 = 2.58.
  • Degrees of freedom = 10 + 10 − 2 = 18. Critical value at p = 0.05: 2.10.
  • 2.58 > 2.10: reject the null hypothesis. Shell height differs significantly between the shores.
  • Why it differs (wave action? predators?) needs a separate hypothesis and investigation.
Quick check
?A student compares the mean lengths of two samples of 8 leaves each and calculates t = 1.92. Part of a table of critical values of t:
Degrees of freedomp = 0.10p = 0.05
71.902.36
141.762.14

Which conclusion is correct?

Mission check

?A student counts the heart rate of Daphnia in pond water (control) and in a caffeine solution, 10 Daphnia in each.
Critical values of t at p = 0.05: 9 degrees of freedom, 2.26; 18, 2.10; 20, 2.09. Which conclusion is correct?
  • Control: mean 212 beats per minute, SD 18.
  • Caffeine: mean 231 beats per minute, SD 20.
Mission 5

Mission 5: Two bars on a graph have different heights. When does that difference actually mean something?

Make a guess: Two bars on a graph have different heights. When does that difference actually mean something?

Mission 5

What an error bar can show

Three kinds of barRoot length / mmn = 10, s = 6.0 mm30405060mean 48544249.946.152.343.7±1 SD±1 SE95%confidence limitsSame data, three different bars:read the legend
  • A bar or point shows a mean. The error bar shows uncertainty or spread.
  • Error bars can show different things. Read the legend first.
  • ±1 SD: the spread of the individual values.
  • ±1 SE or ±2 SE: how precisely the mean is known.
  • 95% confidence limits: the range likely to contain the true mean.
  • Bars from the range (min to max) are sometimes drawn too. They depend on only two values.
Mission 5

Reading SD bars

  • Compare two means, each with ±1 SD bars.
  • Bars overlap widely: many individuals in the two groups are alike. The difference in means may well be due to chance.
  • Bars do not overlap: the difference may be real.
  • SD bars ignore sample size, so they cannot settle it. A statistical test (e.g. a t-test) decides.
  • A small SD means repeats agree closely: a more precise, more repeatable measurement.
Quick check
?Five potato cylinders were bathed in 0.4 mol dm⁻³ sucrose for 24 hours. The mean percentage change in mass was −3.0% with a standard deviation of 1.0%; the mean percentage change in length was −2.0% with a standard deviation of 2.5%. What do these data indicate?
Mission 5

Standard error of the mean

  • Standard error (SE) = s ÷ √n, where s is the SD and n the sample size.
  • SD describes the individuals. SE describes the mean: how close it is likely to be to the true population mean.
  • SE is always smaller than SD (for n > 1).
  • SE falls as n rises: four times as many samples halves the SE.
  • CCEA writes it as σ̂x̄ = √(σ̂² ÷ n): the same thing.
  • Say: "standard error shows how precise the mean is" — not: "standard error is the size of the mistakes".
Mission 5

Worked example: 95% confidence limits

Root lengths of 10 seedlings: x̄ = 48 mm, σ̂ = 6.0 mm.

  • SE = 6.0 ÷ √10 = 1.90 mm.
  • 95% confidence limits = x̄ ± t × SE. Take t from the table at p = 0.05 with n − 1 degrees of freedom.
  • n − 1 = 9: t = 2.262.
  • 48 ± 2.262 × 1.90 = 48 ± 4.3 mm: from 43.7 to 52.3 mm.
  • We are 95% confident the true mean root length lies in this range.
  • Two means whose 95% confidence limits do not overlap are significantly different.
Quick check
?Five potato cylinders were bathed in 0.4 mol dm⁻³ sucrose. Their percentage changes in mass had a mean of −3.0%, a standard deviation of 1.0% and a standard error of 0.45%. What does the standard error indicate?
Mission check

?A student grew bean seedlings with fertiliser X or fertiliser Y, 10 plants each, and plotted mean root length with error bars:
X: mean 42 mm, bar from 39 to 45 mm. Y: mean 50 mm, bar from 47 to 53 mm. Which conclusion is best supported?
  • The bars show 95% confidence limits.
Recap

The big ideas

Mission 1

  • χ² = Σ (O − E)²/E with degrees of freedom = classes − 1; compare with the critical value at p = 0.05.
  • Below it: accept the null hypothesis (difference due to chance). Above it: reject it (significant difference).

Mission 2

  • Mean, median and mode describe the centre; SD describes the typical spread about the mean, using every value.
  • SD is usually more useful than the range, which depends only on the two extreme values.

Mission 3

  • Null hypothesis → test statistic → compare with the critical value at p = 0.05 → significant or not. Neither outcome is proof.
  • χ² compares counts with expected counts; degrees of freedom = categories − 1.
  • A correlation coefficient measures direction and strength (−1 to +1); correlation alone never shows cause.

Mission 4

  • t-test: two means; t = |x̄1 − x̄2| ÷ √(s1²/n1 + s2²/n2); df = n1 + n2 − 2.
  • At or above the critical value at p = 0.05: significant. Never "proved".

Mission 5

  • Read the legend: SD bars show the spread of individuals; SE and confidence bars show how precisely each mean is known.
  • Overlapping SD bars: the difference may be due to chance; a statistical test decides.
  • SE = s ÷ √n; non-overlapping 95% bars (±2 SE or 95% confidence limits) indicate a significant difference.
🏆

Mini-lesson complete!

⭐⭐⭐

You've worked through Statistics and maths skills for CCEA A-level Biology. 🎉

  1. Mission 1Mendel's peas came out at almost three to one, never exactly. How close is close enough?
  2. Mission 2Two sets of results share the same average but look nothing alike. What number captures the difference?
  3. Mission 3A difference between two groups could be real or just luck. How do biologists decide which?
  4. Mission 4Leaves in the shade look bigger than leaves in the sun. How big must the gap be before you believe it?
  5. Mission 5Two bars on a graph have different heights. When does that difference actually mean something?

Your stars: 0 / 0

📣 Smashed it? Share your score

Challenge a friend to beat your stars, or show a parent how you got on.

→ Back to A-Level Biology · CCEA