Lesson 2 of 7 · 12 min

Choosing z for a confidence level

Find the standard normal quantile z that captures the central C% of the distribution, with technology and from the standard values 1.645, 1.96 and 2.576.

What you will be able to do

  • Explain why the central area between −z-z and zz equals the confidence level.
  • Find zz for any confidence level using the inverse normal.
  • Recall zz for 90%, 95% and 99% confidence.

For a large sample, p^\hat{p} is approximately normal with mean pp. So the standardised value

Z=p^−pp(1−p)/nZ = \frac{\hat{p} - p}{\sqrt{p(1-p)/n}}

is approximately standard normal, Z∼N(0,1)Z \sim N(0, 1). If we want an interval that works for 95% of samples, we need the zz value that traps the middle 95% of the standard normal curve.

Interactive

The central 95%

The shaded area between z=−1.96z = -1.96 and z=1.96z = 1.96 is 0.95. Each tail holds 1−0.952=0.025\frac{1 - 0.95}{2} = 0.025.

Key idea

The z for a confidence level C

The confidence level CC is the central area:

P(−z<Z<z)=C.P(-z < Z < z) = C.

The two tails share what is left, so each has area 1−C2\frac{1 - C}{2}, and the area to the left of zz is

P(Z<z)=1+C2.P(Z < z) = \frac{1 + C}{2}.

On a calculator, z=invNorm(1+C2,0,1)z = \text{invNorm}\left(\frac{1+C}{2}, 0, 1\right), or use the inverse normal with a central tail setting.

Finding z for 90% confidence

  1. Find the area in each tail

    1−0.902=0.05\frac{1 - 0.90}{2} = 0.05.

  2. Find the area to the left of z

    0.90+0.05=0.950.90 + 0.05 = 0.95, which is 1+0.902\frac{1 + 0.90}{2}.

  3. Use the inverse normal

    z=invNorm(0.95)≈1.645z = \text{invNorm}(0.95) \approx 1.645.

What area to the left of zz do you enter into the inverse normal to find zz for a 99% confidence interval?

The zz values you will use most. Learn the bold three: the formula book doesn't give them.
Confidence level CCArea in each tailP(Z<z)P(Z < z)zz
80%0.100.901.282
90%0.050.951.645
95%0.0250.9751.96
98%0.010.992.326
99%0.0050.9952.576

Worked example

Find zz for 90% and 85% confidence

  • Simple familiar
  • Technology active

Use technology to find the value of zz used in an approximate confidence interval with confidence level (a) 90% (b) 85%. Give each correct to 3 decimal places.

4 steps

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

Method

  1. Area in each tail: 1−C2\frac{1 - C}{2}.

  2. Area to the left of zz: 1+C2\frac{1 + C}{2}.

  3. z=invNorm(1+C2)z = \text{invNorm}\left(\frac{1+C}{2}\right).

Watch out for

  • Entering CC itself: invNorm(0.85) = 1.036 leaves only 70% in the middle.

Common mistake

invNorm(0.95) is not the 95% value

invNorm(0.95)≈1.645\text{invNorm}(0.95) \approx 1.645 leaves 95% to the left, so only 90% sits in the middle. For 95% confidence you need invNorm(0.975)≈1.96\text{invNorm}(0.975) \approx 1.96. A quick check: a higher confidence level always needs a bigger zz.

Use technology to find zz for a 96% confidence interval. Give your answer correct to 3 decimal places.

Round to 3 decimal places.

Without technology

Recall z=1.645z = 1.645 (90%), 1.961.96 (95%) and 2.5762.576 (99%). The 68–95–99.7% rule says about 95% lies within 2 standard deviations, so 1.96 is "just under 2". Paper 1 questions give zz or use one of these.

With technology

Use the inverse normal with μ=0\mu = 0, σ=1\sigma = 1 and area 1+C2\frac{1 + C}{2} (or choose the central tail option and enter CC). This works for any level, such as 85% or 92%.

Summary

  • The confidence level CC is the central area under the standard normal curve: P(−z<Z<z)=CP(-z < Z < z) = C.

  • Each tail has area 1−C2\frac{1 - C}{2}, so z=invNorm(1+C2)z = \text{invNorm}\left(\frac{1 + C}{2}\right).

  • Standard values: 1.645 for 90%, 1.96 for 95%, 2.576 for 99%.

  • A higher confidence level needs a bigger zz.

Finished the lesson?

Try the 2 quick checks above and the lesson completes itself, or mark it complete now.

Lesson 2 of 70% read

Save your progress with a free account