Lesson 7: Cultural diversity, prejudice and racism
Interpret survey estimates with confidence intervals
SkillsAnalysing cross-cultural dataDiversity, prejudice and racismResearch methods across cultures
Worked example
The Scanlon Foundation Research Institute's 2025 survey estimated the percentage of adults who had experienced discrimination because of their skin colour, ethnic origin or religion in the last 12 months, with 95% confidence intervals: born in Australia 13% [12, 14]; born in the UK, Europe, USA, Canada or NZ 11% [9, 14]; born in Asia or Africa 40% [33, 46].
(a) Calculate the difference, in percentage points, between people born in Asia or Africa and people born in Australia. (b) Use the confidence intervals to decide whether each comparison with the Australian-born group shows a real difference in the population. (c) Suggest why the interval for people born in Asia or Africa is the widest.
Step 1: Calculate the difference
percentage points.
Step 2: Compare Asia or Africa with Australia
The intervals [33, 46] and [12, 14] do not overlap, so the difference is very unlikely to be due to sampling error alone: people born in Asia or Africa are much more likely to report recent discrimination.
Step 3: Compare the UK and Europe group with Australia
The intervals [9, 14] and [12, 14] overlap a lot, so the data do not show a real difference between these groups.
Step 4: Explain the interval widths
Fewer respondents were born in Asia or Africa (1,134, against 5,731 born in Australia). A smaller sample gives a larger standard error, so a wider confidence interval.
: as falls, and the 95% interval grow.
Answer
(a) 27 percentage points. (b) Asia or Africa against Australia: the intervals do not overlap, so the difference is likely real. UK and Europe group against Australia: the intervals overlap, so no clear difference. (c) Fewer respondents in that group, so a larger standard error and a wider interval.
Try a similar one
A fresh question of the same kind, checked as you answer, with hints and a full solution.
Method
Subtract percentages to get a difference in percentage points.
Intervals that do not overlap suggest a real difference; heavily overlapping intervals do not.
Smaller samples give wider intervals.
Watch out for
Writing the difference as 27%. It is 27 percentage points; as a percentage of 13% it would be about 208%.
Concluding the groups are "the same" because intervals overlap. Overlap means the data cannot show a difference, not that there is none.