Lesson 3 of 8 · 16 min

The statistical approach to normality

Use the mean, standard deviation, z-scores and the normal curve to decide whether a score is statistically uncommon, then weigh the statistical approach against the other five.

What you will be able to do

  • Calculate a z-score and use it to decide whether a score is statistically uncommon.
  • Use the 68–95–99.7 rule to estimate the percentage of people above or below a score.
  • Evaluate the statistical approach, including the difference between rare and undesirable.
  • Compare the conclusions of the six approaches for the same behaviour.

Many human characteristics, measured across a large group, form a normal distribution: a symmetrical, bell-shaped curve with most people near the middle and fewer and fewer towards each end. Scores on IQ tests are designed to fall this way, with a mean of 100 and a standard deviation of 15. So are scores on many questionnaires that measure anxiety or mood in large samples.

The statistical approach uses that curve. A score is statistically normal if it lies close to the mean, and statistically abnormal if it lies far out in either tail. The usual cut-off is 2 standard deviations from the mean.

Interactive

Where does "uncommon" start?

The shaded tails are more than 2 standard deviations from the mean: about 2.3% of people in each.

Key idea

z-scores and the 68–95–99.7 rule

A z-score says how many standard deviations a score is from the mean:

z=x−xˉsz = \frac{x - \bar{x}}{s}

A positive zz is above the mean, a negative zz below. In a normal distribution about 68% of scores lie within 1 standard deviation of the mean, 95% within 2 and 99.7% within 3. So about 2.5% of people score more than 2 standard deviations above the mean, and about 2.5% more than 2 below (more precisely, 2.3% each).

Worked example

Use a z-score to judge statistical normality

  • Simple familiar

On a fictional worry questionnaire, scores in a large sample of Australian adults are normally distributed with a mean of 40 and a standard deviation of 8. Jordan scores 58.

(a) Calculate the z-score for Jordan's score.

(b) Using the statistical approach with a cut-off of 2 standard deviations, decide whether Jordan's score is statistically normal.

(c) State one thing the statistical approach cannot tell us about Jordan.

3 steps

Try it yourself first, then check the working one step at a time.

Method

  1. z=x−xˉsz = \frac{x - \bar{x}}{s}, keeping the sign.

  2. Statistically abnormal if ∣z∣>2|z| > 2 (the usual cut-off).

  3. Then say what the statistics cannot show.

Watch out for

  • Dividing the score by the standard deviation (58÷858 \div 8) instead of the difference from the mean.

  • Concluding that Jordan has an anxiety disorder from one score.

On a fictional stress questionnaire, scores in a large sample of adults have a mean of 24 and a standard deviation of 6. Calculate the z-score for a score of 15.

IQ scores have a mean of 100 and a standard deviation of 15. Using the 68–95–99.7 rule, about what percentage of people score above 130?

Choose one answer

Common mistake

Rare is not the same as a problem

A person with an IQ of 145 is as statistically abnormal as a person with an IQ of 55: each is 3 standard deviations from the mean. The statistical approach cannot tell desirable rarity from undesirable rarity. It also says nothing about behaviour that is common but harmful: in a bad flu season, being unwell is common. To decide whether rare matters, you need another approach, usually the functional one.

The statistical approach in diagnosis. Clinicians do use statistics, but never on their own. For example, DSM-5-TR's criteria for an intellectual developmental disorder include intellectual functioning about 2 standard deviations or more below the mean (around an IQ of 70, allowing for measurement error), and deficits in adaptive functioning in everyday life, beginning in childhood. A low score alone is not enough: the functional approach has to agree.

Strengths: it is objective, it uses data rather than opinion, and the same cut-off can be applied to anything that can be measured. Limitations: it does not distinguish desirable from undesirable rarity; the cut-off (2 standard deviations) is a convention, not a natural boundary; it ignores culture and situation (what is rare in one population may be common in another); and many behaviours are not measured on a scale at all.

Statistically abnormal, not a problem

An IQ score of 140. Running a marathon in under 3 hours. Being able to name every Premier of Queensland. Each is rare, and none gets in the way of the person's life.

Statistically normal, still a problem

Feeling persistently low after a long illness, in a year when many people have been ill. Common does not make it harmless: the functional approach shows the cost.

Interactive

Shade the uncommon scores

A fictional worry questionnaire with mean 40 and standard deviation 8. Scores above 56 are more than 2 standard deviations above the mean.

Select all the statements that are limitations of the statistical approach to normality.

Select all that apply.

Worked example

Rare in both directions on the IQ scale

  • Complex familiar

IQ scores have a mean of 100 and a standard deviation of 15. Two fictional adults are assessed: Ella scores 145 and Sam scores 55.

(a) Use the 68–95–99.7 rule to estimate the percentage of people who score above 145.

(b) Explain why the statistical approach judges both scores the same way.

(c) Use the functional approach to explain why the two scores may have very different meanings.

4 steps

Try it yourself first, then check the working one step at a time.

Method

  1. Convert each score to a z-score.

  2. Use 68–95–99.7 and split the outside percentage between the two tails.

  3. Bring in another approach to judge what the rarity means.

Watch out for

  • Giving 0.3% for one tail: that is both tails together.

  • Assuming a low IQ score alone means a disorder; adaptive functioning must also be assessed.

Exam tip

Show the z-score

When you use the statistical approach on data, show the calculation (z=x−xˉsz = \frac{x - \bar{x}}{s} with the numbers in), state the cut-off you are using ("more than 2 standard deviations from the mean"), and then interpret: "so the score is statistically uncommon". Then, if the question asks you to evaluate, say what the statistics cannot tell you.

Summary

  • The statistical approach calls a score abnormal when it is uncommon, usually more than 2 standard deviations from the mean.

  • z=x−xˉsz = \frac{x - \bar{x}}{s}; about 68%, 95% and 99.7% of scores lie within 1, 2 and 3 standard deviations of the mean.

  • About 2.5% (more precisely 2.3%) of people lie more than 2 standard deviations above the mean, and the same below.

  • Strengths: objective, based on data. Limitations: rare is not undesirable, the cut-off is a convention, and it ignores culture and situation.

  • Clinicians combine statistics with other approaches: a low IQ score must come with difficulties in adaptive functioning for an intellectual developmental disorder.

Finished the lesson?

Try the 3 quick checks above and the lesson completes itself, or mark it complete now.

Lesson 3 of 80% read

Save your progress with a free account