The Chromozone / IAL Biology / Unit 4 / Topic 5

Directed study: exam technique

Exam technique: statistics in the exam

Spec 5.14Directed study: exam techniqueExam technique: Statistics in the examAbout 40 minutes
Back to the lesson

Directed study: exam technique. About 40 minutes, independent study, 23 marks. One skill: choosing the right statistical test, writing the null hypothesis, finding the degrees of freedom and the critical value, and writing the conclusion in words. Type your answers in the boxes, then hand them in at the bottom of the page. Steps 1 and 2 stay open to use; the mark schemes open after you hand in: then mark your work point by point and correct it in a different colour. Teachers: set this after the lesson Core practical 11: sampling methods and analysing the results (lesson 2 of 2).

Teach these first

This page builds on these lessons. Teach them before students start here:

How this page works

One exam technique, in three steps: learn how to do it, see it done with the marks shown, then do it yourself on this lesson's content.

Step 1: how to do it 6 min

Use these steps whenever a question gives you data from a habitat study and asks about a statistical test. The calculation is in the three statistics lessons. This page trains the choices and the sentences around it.

  1. Ask what you measured. Two measured variables, and you want a link between them: Spearman's rank.
  2. Two groups, a measurement from each, and you want to compare the means: Student's t-test.
  3. Counts in categories, compared with numbers you expect (a ratio or an equal spread): chi-squared.
  4. Write the null hypothesis: "no significant" correlation or difference, naming what you measured.
  5. Find the number to look up. Spearman's rank: n, the pairs. t-test: n1 + n2 − 2. Chi-squared: categories − 1.
  6. Read the critical value at p = 0.05. State your value and the critical value. Ignore a minus sign.
  7. Equal to or greater than the critical value: reject the null hypothesis. Smaller: accept it.
  8. Write the conclusion in words, about the biology, and give the direction.
Choosing the test and writing the null hypothesis
TestUse it whenNull hypothesis (H0)Number to look up
Spearman's rankYou have two measured variables and look for a correlation.There is no significant correlation between (variable 1) and (variable 2).n, the number of pairs
Student's t-testYou compare the means of two groups of measurements.There is no significant difference between the mean (measurement) of (group 1) and (group 2).n1 + n2 − 2
Chi-squaredYou compare counts in categories with expected numbers.There is no significant difference between the observed and the expected numbers (say what you expect).Categories − 1

How the mark scheme rewards this. Pearson gives a mark for the critical value, a mark for comparing it with the calculated value and deciding to reject or accept the null hypothesis, and a mark for what the result means for the biology. Some questions add a mark for a comment on the variability of the data, such as range bars that overlap. "It is significant" on its own, or "rs is close to 1", is not a decision: only the comparison with the critical value decides.

Two true points that earn no mark in a conclusion. "The probability that the result is due to chance is less than 5%" is what p = 0.05 means. "A correlation does not show cause" is also true. Write them when a question asks what p = 0.05 means, or asks you to evaluate a conclusion.

Write the sentence in this order. "The calculated value, (value), is greater than the critical value, (value), for (n or df) at p = 0.05, so I reject the null hypothesis. There is a significant (correlation or difference) between (what you measured)." Then give the direction. If the value is smaller, write "less than", "accept the null hypothesis" and "no significant". Then any difference is due to chance.

Common misconceptions and exam errors

These come up again and again in the reports on real exam answers. Each one costs marks.

  1. Naming a t-test or chi-squared test for a link between two measured variables. Pearson does not accept another test when a correlation test is wanted. Two measured variables and a link: Spearman's rank. A t-test needs two groups. Chi-squared needs counts in categories.
  2. Comparing with "the critical value" without giving the numbers. One mark scheme does not allow "t is less than the critical value" when the value of t is not given. Write both numbers, your calculated value and the critical value, then reject or accept.
  3. Using the critical value for the wrong number of pairs or degrees of freedom. One mark scheme rejects every critical value except the right one. Use n for Spearman's rank, n1 + n2 − 2 for the t-test and categories − 1 for chi-squared. Then read the p = 0.05 column, unless the question names another.
  4. Writing a negative value of t in the comparison. One mark scheme does not allow a negative t when it is compared with the critical value. Ignore the sign and compare the size: "2.52 is greater than 2.093". The sign only tells you which group had the larger mean, or that a correlation is negative. Say that in words.
  5. Writing that one variable "causes" the other after a significant correlation. Pearson does not accept a causal relationship. Write that there is a correlation, and say that another factor may change with both variables.

Step 2: a worked example 5 min

This example is for learning. Its 6 marks do not count in your total.

Question: A student studied a bed of seagrass in a shallow lagoon. She laid a transect from the shore out into the lagoon. At nine quadrats she measured the depth of the water (cm) and counted the seagrass shoots. She calculated rs = −0.82.

Critical values of rs at p = 0.05
Number of pairs (n)Critical value
70.786
80.738
90.700
100.648

(a) Name the statistical test she should use. Give a reason. (2)
(b) State a suitable null hypothesis. (1)
(c) Use the table to give a conclusion. (3)

Answer 1 3 of 6 marks

(a) Spearman’s rank (1), because she wants to see if depth and shoots are linked (1). (b) The null hypothesis is that deeper water gives fewer shoots. (c) The calculated value is −0.82. The critical value is 0.700 (1). −0.82 is less than 0.700, so I accept the null hypothesis. Deeper water causes fewer shoots.

Why: The test and the reason earn two marks, and the critical value earns one. The null hypothesis is wrong: it states the prediction, and a null hypothesis says there is no significant correlation. The sign was not ignored, so the decision is wrong (0.82 is greater than 0.700), and the meaning that follows is lost too. "Causes" is never credited for a correlation.

Answer 2 6 of 6 marks

(a) Spearman’s rank correlation (1), because she has two measured variables from the same quadrats and wants to know if they are correlated (1). (b) There is no significant correlation between the depth of the water and the number of seagrass shoots (1). (c) For 9 pairs the critical value at p = 0.05 is 0.700 (1). Ignoring the sign, 0.82 is greater than 0.700, so I reject the null hypothesis (1). There is a significant negative correlation: there are fewer shoots in deeper water (1). The correlation does not show that depth itself is the cause: another factor, such as the light reaching the seagrass, falls with depth.

Why: Each mark is a separate point: the test, the reason, a null hypothesis in the right form, the critical value, both numbers compared with the decision, and the meaning for the biology with the direction. She compared the size of rs and gave the direction in words. The last sentence is true and helps in an evaluate question, but it earns no mark in a conclusion.

Step 3: your turn 25 min

Use the steps in Step 1. Part 1 is guided. In parts 5 and 6 the calculated value is given, so you find the degrees of freedom and the critical value, and decide. Open the mark scheme only after you hand in.

  1. Context for part 1. A student laid ten quadrats up a slope from the edge of a lake. In each quadrat she measured the percentage of water in the soil and the percentage cover of sedge, a grass-like plant. She wants to know if the two are linked.

    Name the statistical test she should use and give a reason. State a null hypothesis. State the number you use to find the critical value in the table. (4)

    Test
    Reason
    Null hypothesis
    Number to look up
  2. Context for part 2. A student measured the length of 12 fronds of one fern species at the edge of a forest and 12 fronds of the same species inside the forest. She wants to know if the mean length differs.

    Name the statistical test she should use and give a reason. State a null hypothesis. Give the degrees of freedom. (4)

  3. Context for part 3. A student counted the webs of one species of spider on the posts of a long wooden fence. For each web she recorded the side of the post it was on: north, east, south or west. She found 72 webs. If the spiders had no preference, there would be the same number of webs on each side.

    Name the statistical test she should use and give a reason. State a null hypothesis. Give the degrees of freedom. (4)

  4. Context for part 4. A student counted the ants in a quadrat at 10 random places on the edge of a paddy field and at 10 random places in the nearby forest. She says: "I counted things, so I will use a chi-squared test."

    Name the correct statistical test. Explain why a chi-squared test is not suitable. State a null hypothesis. (3)

  5. Tables for parts 5 and 6. Both tables give the critical values at p = 0.05.

    Critical values of χ2 at p = 0.05
    Degrees of freedomCritical value
    13.84
    25.99
    37.81
    49.49
    511.07
    612.59
    Critical values of t at p = 0.05
    Degrees of freedomCritical value
    122.179
    142.145
    162.120
    182.101
    192.093
    202.086
    222.074

    Context for part 5. A student counted 100 dragonflies perched on five kinds of plant beside a pond. The null hypothesis is that the dragonflies are spread equally, 20 on each kind of plant. She calculated χ2 = 11.3.

    Give the degrees of freedom and the critical value. Use them to decide whether to accept or reject the null hypothesis, and write a conclusion about the dragonflies. (4)

  6. Context for part 6. A student measured the width of 10 leaves from a shrub in full sun and 11 leaves from the same species in deep shade. The mean width was 4.1 cm in the sun and 5.0 cm in the shade. She calculated t = −2.52.

    Give the degrees of freedom and the critical value. Use them to write a conclusion about the width of the leaves. (4)

Mark scheme

1 Choose: a link between two variables (4)

  • Spearman's rank (correlation coefficient) (1)
  • Two measured variables, from the same quadrats / looking for a link (correlation) between them (1)
  • There is no (significant) correlation between the percentage of water in the soil and the percentage cover of sedge (1)
  • n, the number of pairs of values = 10 (1)
  • DO NOT ACCEPT a t-test or chi-squared test
  • ACCEPT "no relationship / association" for "no correlation"
  • DO NOT ACCEPT a null hypothesis that says there is a correlation, or that one variable affects the other
  • In this course the Spearman's rank table is read with n, so DO NOT ACCEPT degrees of freedom or 8 or 9

2 Choose: two groups (4)

  • Student's t-test (1)
  • It compares the means of two groups of measurements (the lengths of fronds at the edge and inside the forest) (1)
  • There is no (significant) difference between the mean length of the fern fronds at the forest edge and inside the forest (1)
  • df = 12 + 12 − 2 = 22 (1)
  • DO NOT ACCEPT chi-squared (lengths are measurements, not counts in categories) or Spearman's rank (there is no pair of variables)
  • The null hypothesis must name what was measured and both groups

3 Choose: counts in categories (4)

  • Chi-squared test (1)
  • The data are counts in categories (the four sides of the posts), compared with expected numbers (18 on each side) (1)
  • There is no (significant) difference between the observed numbers of webs on each side and the expected numbers / the spiders show no preference for any side (18 on each) (1)
  • df = 4 − 1 = 3 (1)
  • DO NOT ACCEPT a t-test (there are no two groups of measurements)
  • DO NOT ACCEPT df = 71 (the number of webs minus 1) or df = 4

4 Counts that are not categories (3)

  • Student's t-test (1)
  • Each quadrat gives one number, and the means of two groups (paddy edge and forest) are compared. Chi-squared compares counts in categories with expected numbers (1)
  • There is no (significant) difference between the mean number of ants per quadrat at the paddy field edge and in the forest (1)
  • The reason must say why chi-squared does not fit: these are counts per quadrat, compared as the means of two groups, not counts in categories against expected numbers
  • DO NOT ACCEPT Spearman's rank: there is no pair of variables for each quadrat

5 Decide: chi-squared (4)

  • df = 5 − 1 = 4 (1)
  • Critical value = 9.49 (at p = 0.05) (1)
  • 11.3 is greater than 9.49, so reject the null hypothesis (1)
  • There is a significant difference: the dragonflies are not spread equally between the plants / they prefer some kinds of plant (1)
  • Both values must be stated for the third point
  • ALLOW ecf: a correct comparison with a wrong critical value they chose. No mark for the value
  • IGNORE the chance statement (less than 5%): it is true but not needed
  • DO NOT ACCEPT "this proves": a test gives a probability, not proof

6 Decide: t-test (4)

  • df = 10 + 11 − 2 = 19 (1)
  • Critical value = 2.093 (at p = 0.05) (1)
  • Ignoring the sign, 2.52 is greater than 2.093, so reject the null hypothesis (1)
  • There is a significant difference between the mean leaf width in the sun and in the shade (the leaves are wider in the shade) (1)
  • DO NOT ACCEPT "−2.52 is less than 2.093" or a negative t in the comparison
  • ALLOW ecf: a comparison made with a critical value that follows from an incorrect df
  • IGNORE the chance statement: it is true but earns no mark

Self-marking checklist 4 min

Before you hand in, check that:

  • you named the test from the data: two measured variables (Spearman's rank), two groups of measurements (t-test) or counts in categories (chi-squared), and gave that as your reason;
  • every null hypothesis starts "no significant" and names what was measured;
  • every decision gives the calculated value and the critical value, ignores any minus sign, and says reject or accept;
  • every conclusion says what the result means for the biology, with the direction, and you did not write that one variable causes the other.

After you hand in, the mark schemes open. Then:

  • tick each mark point you earned: a decision without both values earns nothing;
  • for each point you missed, rewrite it in a different colour in the order of the sentence in Step 1;
  • write your total out of 23, and one target, for example: "I will always write both numbers and then reject or accept."

Back to the lesson

Was this page useful?

Goes straight to the two people who make these pages. Please don't include students' names or other personal details.