ASCENTLEARNING
Join a weekly class
Skip to learning content
Year 11 · Statistics · 18 min

Outliers and fair conclusions

Investigate unusual observations without hiding genuine extremes.

01 / LEARN

The idea, step by step

A flag starts an investigation

One common rule flags values below Q1−1.5⁢I⁢Q⁢R or above Q3+1.5⁢I⁢Q⁢R. Use the stated quartile convention. A flagged value is distant from the middle, but the rule does not prove an error.

Check original records, units and population definitions. A 160-minute trip could be genuine; 1600 might be a transcription error. Evidence, not a preferred mean, should guide treatment.

Show sensitivity honestly

Correct and document verified errors. Explain any defensible pre-established exclusion rule. Otherwise retain relevant genuine observations and choose suitable summaries, or show a sensitivity analysis with and without the extreme.

Extremes can strongly change mean and range while leaving median/IQR relatively stable. They may reveal a real subgroup or important event. Removing them silently can misrepresent the group.

A common trap

Delete everything outside the fences.

Instead: A fence flags an observation to investigate, not a mandatory deletion.

02 / WORKED EXAMPLES

See the thinking

Example 1

Find fences

Q1=10, Q3=18. Test−3 and 31.

  1. IQR 8.
  2. Fences 10−12=−2 and 18+12=30.
  3. −3 and 31 are outside.
Both flagged; neither automatically wrong.
Example 2

Sensitivity

Compare 4, 5, 5, 6, 30 with and without 30.

  1. With all: mean 10, median 5.
  2. Without 30: mean 5, median 5.
  3. Mean changes greatly; median does not.
Exclusion changes mean 10 to 5 but still needs a valid reason.

Check your understanding

Q1=6, Q3=10. Upper fence?

Show answer

16.

IQR 4; 10+1.5⁢(4)=16.

03 / WORKSHEET

Your turn

6 questions · 12 marks · Work at your own pace.

Read the questions online
  1. Find fences for Q1=12, Q3=20.
    2 marks
  2. A value equals the upper fence 32. Is it flagged by the strict outside rule?
    2 marks
  3. Source confirms 1.75 m entered as 1.75 cm. What correction?
    2 marks
  4. Find mean and median of 3, 4, 4, 5, 24.
    2 marks
  5. Slow runners are removed to improve mean race time. Explain the problem.
    2 marks
  6. Results change greatly without one genuine extreme. What should be reported?
    2 marks
Read the worked solutions

Equivalent correct methods earn the corresponding marks. Include units where asked.

  1. Find fences for Q1=12, Q3=20.
    2 marks

    IQR 8; subtract/add 12.

    0 and 32.
  2. A value equals the upper fence 32. Is it flagged by the strict outside rule?
    2 marks

    The rule uses values above, not equal to, the fence.

    No.
  3. Source confirms 1.75 m entered as 1.75 cm. What correction?
    2 marks

    It is a verified unit mismatch.

    175 cm, documenting the unit correction.
  4. Find mean and median of 3, 4, 4, 5, 24.
    2 marks

    Total 40 divided by count 5; third value 4.

    Mean 8; median 4.
  5. Slow runners are removed to improve mean race time. Explain the problem.
    2 marks

    Exclusion cannot be justified by wanting a nicer result.

    Selective removal of genuine outcomes biases the description.
  6. Results change greatly without one genuine extreme. What should be reported?
    2 marks

    Readers need to see the sensitivity.

    Both relevant results, the treatment rule and a cautious conclusion.
Checking saved progress…

Continue learning

Communicating findings and limitations

Next lesson
Curriculum reference & lesson scope

SACE Board of South Australia

Stage 1 General Mathematics: statistical investigation, univariate data, sampling, displays, summaries and interpretation.

Official curriculum source ↗
Back to Statistics

Save your next step.

Create a free account to save your learning progress. Lessons and downloads stay free, with or without an account.

Create free account