You are on page 1of 2

Inverse Normal Distribution [TI-nSpire]

Inverse Normal Distribution


A reverse process for finding values when given the probability for a normal distribution
Probabilities may be given as areas, percentiles (𝑥th percentile = ⎯⎯⎯), or quantiles (a decimal).
The TI-nSpire can only do inverse normal for Pr(𝑋 < 𝑥) = 𝑎.
Therefore, if it is not in that form, it must be rewritten so that it is.

b → 5 (𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦) → 5 (𝐷𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛𝑠) → 3 (𝐼𝑛𝑣𝑒𝑟𝑠𝑒 𝑁𝑜𝑟𝑚𝑎𝑙 … )


b → 6 (𝑆𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐𝑠) → 5 (𝐷𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛𝑠) → 3 (𝐼𝑛𝑣𝑒𝑟𝑠𝑒 𝑁𝑜𝑟𝑚𝑎𝑙 … )
Pr(𝑋 < 𝑥) = area invNorm(area, 𝜇, 𝜎)
Pr(𝑋 < 𝑥) = 𝑝 𝑥 = invNorm(𝑝, 𝜇, 𝜎)
Pr(𝑍 < 𝑧) = 𝑝 z = invNorm(𝑝, 0,1)
Pr(𝑋 > 𝑥) = 𝑝 𝑥 = 2𝜇 − invNorm(1 − 𝑝, 𝜇, 𝜎)
Pr(𝑍 > 𝑧) = 𝑝 z = invNorm(1 − 𝑝, 0,1)
Pr(𝜇 − 𝑥 < 𝑋 < 𝜇 + 𝑥) = 𝑝 1−𝑝
𝑥 = invNorm 1 − ⎯⎯⎯⎯⎯, 𝜇, 𝜎 − 𝜇
2
Pr(−𝑧 < 𝑍 < 𝑧) = 𝑝 1−𝑝
𝑧 = invNorm 1 − ⎯⎯⎯⎯⎯, 0,1
2

Example Example Example


Pr(𝑍 < 𝑐) = 0.29 Pr(𝑍 ≥ 𝑐) = 0.19 Pr(−𝑐 < 𝑍 < 𝑐) = 0.86
𝑐 = invNorm(0.29,1,0) 𝑐 = invNorm(1 − 0.19,0,1) 1 − 0.86
𝑐 = −0.5534 𝑐 = invNorm(0.81,0,1) 𝑐 = invNorm 1 − ⎯⎯⎯⎯⎯⎯⎯⎯,0,1
2
𝑐 = 0.8779 c = invNorm(0.93,0,1)
𝑐 = 1.4758
Example VCAA 2000 Exam 2 Question 2a
Pedro fishes in an area where there are large numbers of salmon. It is known that the lengths of salmon in this
area are normally distributed with mean 36.0 cm and standard deviation 2.5 cm. Assume that all salmon are
equally likely to be caught. Salmon with lengths in the top 15% of lengths of salmon in the area are called
gourmet salmon. The minimum length of a gourmet salmon is

𝑋~N 36, 2.5 , Pr(𝑋 > 𝑔) = 15% ⇒ Pr(𝑋 < 𝑔) = 85%


On CAS: Area = 0.85, 𝜇 = 36, 𝜎 = 2.5 ⇒ invNorm(0.85,36,2.5) = 38.59108
∴ 𝑔 = 38.6 cm

Finding 𝜇 and 𝜎
Finding 𝜇 and 𝜎
𝑥−𝜇
1. Standardise given 𝑥 probabilities using 𝑧 = ⎯⎯⎯⎯⎯ in terms of 𝜇 and 𝜎
𝜎
𝑥−𝜇
Pr(𝑋 < 𝑥) = 𝑎 ⟹ Pr 𝑍 < ⎯⎯⎯⎯⎯ = 𝑎
𝜎
2. Use invNorm(𝑎, 0,1) to find the corresponding 𝑧 values for the given 𝑥 values
𝑥−𝜇
𝑧 = ⎯⎯⎯⎯⎯= invNorm(𝑎, 0,1)
𝜎
3. Solve the two equations simultaneously

Example VCAA 2002 Exam 2 Question 2b


Emmy is gathering data on two particular species of yellow butterflies, Fhaisi and Jojo, which are very difficult
to tell apart. Both species are equally likely to be caught. One technique for telling the difference between the
two species is by measuring the length of their antennas. 8% of Jojo butterfly antennas are shorter than 19
mm and 8% of Jojo butterfly antennas are longer than 28 mm. Assume that the antenna lengths of Jojo
butterflies are also normally distributed.
The mean and the standard deviation of antenna lengths of Jojo butterflies are

𝑋~N 𝜇, σ , 𝑍~N(0,1)

Using Simultaneous Equations Using Symmetry


19 − 𝜇 Since Pr(𝑋 < 19) = Pr(𝑋 > 28), 19 and 28
Pr(𝑋 < 19) = Pr 𝑍 < ⎯⎯⎯⎯⎯⎯ = 8% they are the same distance from the mean.
𝜎
On CAS: Area = 0.08, 𝜇 = 0, 𝜎 = 1 28 + 19
∴ 𝜇 = ⎯⎯⎯⎯⎯⎯⎯= 23.5 mm
⇒ invNorm(0.08,0,1) = −1.40507 2

28 − 𝜇 −4.5
Pr(𝑋 > 28) = Pr 𝑍 > ⎯⎯⎯⎯⎯⎯ = 8% Pr(𝑋 < 19) = Pr 𝑍 < ⎯⎯⎯⎯ = 8%
𝜎 𝜎
28 − 𝜇 On CAS: Area = 0.08, 𝜇 = 0, 𝜎 = 1
⇒ Pr 𝑍 < ⎯⎯⎯⎯⎯⎯ = 92% ⇒ invNorm(0.08,0,1) = −1.4050
𝜎
On CAS: Area = 0.92, 𝜇 = 0, 𝜎 = 1 −4.5
⇒ invNorm(0.92,0,1) = 1.40507 ⎯⎯⎯⎯ = −1.40507
𝜎
⇒ 4.5 = 1.40507𝜎
19 − 𝜇 28 − 𝜇 ⇒ 𝜎 = 3.20268 ≈ 3.2 mm
∴ ⎯⎯⎯⎯⎯⎯= −1.40507 and ⎯⎯⎯⎯⎯⎯= 1.40507
𝜎 𝜎

Solving simultaneously on CAS:


𝜇 = 23.5 mm, 𝜎 = 3.20268 ≈ 3.2 mm

You might also like