4.12 Standardisation and inverse normal with unknown parameters

4.12 Standardisation and inverse normal with unknown parameters

IB Maths: Analysis and Approaches SL: Revision notes

Studying this course? Keep your progress in one place.

Key facts

  • Standardise: z=x−μσz=\dfrac{x-\mu}{\sigma}, the number of standard deviations from the mean; x=μ+zσx=\mu+z\sigma.
  • P(X<x)=P(Z<x−μσ)P(X<x)=P\left(Z<\frac{x-\mu}{\sigma}\right) where Z∼N(0,1)Z\sim N(0,1).
  • Compare values from different distributions using their zz-values.
  • If μ\mu or σ\sigma is unknown, find zz from N(0,1)N(0,1) then solve x−μσ=z\frac{x-\mu}{\sigma}=z.
  • Two conditions give simultaneous equations for μ\mu and σ\sigma.

Standardising

The zz-value counts how many standard deviations xx lies from the mean, turning any normal variable into the standard normal Z∼N(0,1)Z\sim N(0,1).

For X∼N(μ,σ2)X\sim N(\mu,\sigma^2), z=x−μσz=\frac{x-\mu}{\sigma}. A positive zz is above the mean and a negative zz is below it. Rearranged, x=μ+zσx=\mu+z\sigma.

So P(X<x)=P(Z<x−μσ)P(X<x)=P\left(Z<\frac{x-\mu}{\sigma}\right). Note that N(50,16)N(50,16) has σ=4\sigma=4, not 16.

−4−3−2−112340.10.20.30.4xyz = 1.5standard normal
For X∼N(50,42)X\sim N(50,4^2), P(X>56)P(X>56) is the shaded area P(Z>1.5)P(Z>1.5) under the standard normal curve.

Worked example

X∼N(50,42)X\sim N(50,4^2). Find P(X>56)P(X>56).

For N(50,16)N(50,16), what is σ\sigma in z=x−μσz=\frac{x-\mu}{\sigma}?

Comparing different distributions

Compare zz-values, not raw marks: the larger zz is further above its own group's mean, relative to the spread.

Ben's z=2.2z=2.2 beats Amira's z=1.5z=1.5, so Ben did relatively better even though both marks are above average.

To find an equivalent value in another distribution, keep the zz-value and use x=μ+zσx=\mu+z\sigma with the new parameters: 70+1.5×5=77.570+1.5\times5=77.5.

304050607080901000.020.040.060.08xyAmira: 72, z = 1.5Ben: 81, z = 2.2Amira: N(60, 8²)Ben: N(70, 5²)
Amira (72 in N(60, 8²)) and Ben (81 in N(70, 5²)). Ben is 2.2 standard deviations above his mean, Amira only 1.5, so Ben did relatively better.

Amira

Mark:
72
Mean μ\mu:
60
Standard deviation σ\sigma:
8
zz-value:
1.5

Ben

Mark:
81
Mean μ\mu:
70
Standard deviation σ\sigma:
5
zz-value:
2.2

Cara scores 65 where μ=50\mu=50, σ=10\sigma=10. Dev scores 90 where μ=70\mu=70, σ=20\sigma=20. Who did relatively better?

One unknown parameter

Use inverse normal on the standard normal ZZ to find zz, then solve x−μσ=z\frac{x-\mu}{\sigma}=z for the unknown.

For T∼N(μ,62)T\sim N(\mu,6^2) with P(T<40)=0.2P(T<40)=0.2, inverse normal on Z∼N(0,1)Z\sim N(0,1) with area 0.2 gives z=−0.8416z=-0.8416. Then 40−μ6=−0.8416\frac{40-\mu}{6}=-0.8416 gives μ=45.0\mu=45.0.

For an upper tail, convert first: P(X>a)=0.1P(X>a)=0.1 means P(X<a)=0.9P(X<a)=0.9, so z=1.2816z=1.2816.

−4−3−2−112340.10.20.30.4xyz = −0.8416standard normal
P(Z<z)=0.2P(Z<z)=0.2: the area to the left of z=−0.8416z=-0.8416 is 0.2, so zz is negative.

Worked example

T∼N(μ,62)T\sim N(\mu,6^2) and P(T<40)=0.2P(T<40)=0.2. Find μ\mu.

P(X>a)=0.1P(X>a)=0.1 for X∼N(μ,σ2)X\sim N(\mu,\sigma^2). What area to the left of aa do you use in inverse normal?

Both parameters unknown

Two probability conditions give two equations in μ\mu and σ\sigma. Find a zz-value for each, then solve them simultaneously.

For bolts with 5% shorter than 49.2 mm and 10% longer than 50.9 mm: μ−1.6449σ=49.2\mu-1.6449\sigma=49.2 and μ+1.2816σ=50.9\mu+1.2816\sigma=50.9. Solving gives σ=0.581\sigma=0.581 and μ=50.2\mu=50.2.

For a symmetric condition such as "3% differ from μ\mu by more than dd", put half in each tail: P(Z>dσ)=0.015P(Z>\frac{d}{\sigma})=0.015.

  1. 1

    Write each condition

    e.g. P(X<49.2)=0.05P(X<49.2)=0.05 and P(X>50.9)=0.1P(X>50.9)=0.1

  2. 2

    Find each zz on N(0,1)N(0,1)

    z=−1.6449z=-1.6449 and z=1.2816z=1.2816

  3. 3

    Form two equations

    49.2−μσ=−1.6449\frac{49.2-\mu}{\sigma}=-1.6449, 50.9−μσ=1.2816\frac{50.9-\mu}{\sigma}=1.2816

  4. 4

    Solve simultaneously

    σ=0.581\sigma=0.581, μ=50.2\mu=50.2

Solving for two unknowns

"4% of values lie more than dd from the mean, either side." What tail area goes on each side?

Try an exam question

X∼N(μ,σ2)X\sim N(\mu,\sigma^2) with P(X<30)=0.2P(X<30)=0.2 and P(X>45)=0.1P(X>45)=0.1. Find μ\mu and σ\sigma.

[4 marks]

That's the notes covered.

Carry on to the next subtopic.

Try a question

Find the zz-value of x=56x=56.

1 mark, marked against the mark scheme

Revision notes