Every 9700 Paper 5 asks you to process someone else’s results. Work through two investigations with fresh data each time: find the mean, standard deviation (with n − 1) and standard error of a sample, draw the bar chart and its 95% CI error bars on the grid by hand, then carry out a t-test, and after that a chi-squared test on a genetic cross or Spearman’s rank correlation along a transect. State each null hypothesis, calculate the statistic to the right number of significant figures, work out the degrees of freedom, read the critical value at p = 0.05 from the table and write the conclusion the way the mark scheme does. Finish by evaluating the method. The formulae and critical values are the ones printed in the syllabus. Allow about 35 minutes.
This workout needs JavaScript.
The mean to one decimal place, the standard deviation using n − 1 and given to the same number of decimal places as the table (9700 November 2023 examiner report: s = 3 instead of 3.2 was not credited), and SE = s / √n for both samples.
Each bar top within half a small square of its mean, and each error bar running from mean − 2SE to mean + 2SE. Only standard error or 95% CI bars allow a judgement about significance; standard deviation bars do not (November 2023 report).
A null hypothesis of no difference between the two means, t to three significant figures (1.883, not 1.9: June 2024 report), degrees of freedom n₁ + n₂ − 2 with the matching critical value at p = 0.05, then the comparison and the conclusion: “there is a significant difference ... the null hypothesis is rejected” or “no significant difference ... accepted”.
A variable that really affects the measurement and was not standardised, and a change that reduces the standard error: more measurements from more sites, not a finer ruler or a different probability level.
Chi-squared: expected numbers (not ratios), each (O − E)²/E, χ² to three significant figures, degrees of freedom = classes − 1, and accept or reject correctly (June 2023 report: many rejected a null hypothesis their value supported). Spearman’s rank: ranks and D², rs to three significant figures, the critical value for the number of pairs, and a conclusion that names a significant positive or negative correlation, never a “significant difference”.
The marking follows the conventions in published Cambridge mark schemes. The experiments, readings and wording on this page are our own.