S1: Chapter 9 The Normal Distribution Dr J Frost (jfrost@tiffin.kingston.sch.uk) www.drfrostmaths.com Last modified: 7th February 2016 Teacher Guidance Possible lesson structure: Lesson 1: Introduction and using z-tables Go > Lesson 2: Standardising, calculating probability of ranges Lesson 3: Reverse z-table, calculating values given probabilities Lesson 4: More on reverse z-tables: Quartiles, dealing with π(π < π < π) and π π − π < π < π + π Lesson 5 : Unknown π and/or π Lesson 6: Harder questions (including conditional probabilities) Go > Go > Go > What does it look like? The following shows what the probability distribution might look like for a random variable X, if X is the height of a randomly chosen person. We expect this ‘bell-curve’ shape, where we’re most likely to pick someone with a height around the mean of 180cm, with the probability diminishing symmetrically either side of the mean. p(x) 180cm A variable with this kind of distribution is said to have a normal distribution. Height in cm (x) For normal distributions we tend to draw the π¦ axis at the mean for symmetry. π(π₯) What does it look like? We can set the mean π and the standard deviation π of the Normal Distribution. Normal Distribution Q & A Q1 For a Normal Distribution to be used, the variable has to be: continuous ? Q2 With a discrete variable, all the probabilities had to add up to 1. For a continuous variable, similarly: the area under the probability graph has ? to be 1. Q3 To find π 170 < π < 190 , we could: find the area between these values. π π₯ Q4 ? Would we ever want to find π π = 200 say? Since height is continuous, the probability someone is ‘exactly’ 200cm is infinitesimally small. So not a ‘probability’ in the normal sense. ? 170cm 180cm 190cm Height in cm (π₯) Notation If a variable π is ‘normally distributed’ (i.e. its probability function uses a normal distribution), then we write: ...is distributed... ...using a Normal distribution with mean π and variance π 2 The random variable X... π~π π, π 2 Example: “π represents the height of a randomly chosen person, with mean 160cm and standard deviation 10cm.” 2 π~π 160,10 ? Z value ! The Z value is the number of standard deviations a value is above the mean. Example p(x) IQ, by definition, is normally distributed for a given population. By definition, π = 100 and π = 15 i.e. π~π 100,152 IQ Z 100 0 130 85 165 100 IQ (π₯) 62.5 ? 2 ? -1 ? 4.333 ? -2.5 ? Z table Minimise A z-table allows us to find the probability that the outcome will be less than a particular z value. For IQ, π·(π < π) would mean “the probability your IQ is less than 130”. (You can find these values at the back of your textbook, and in a formula booklet in the exam.) Expand π π < 2 = 0.9772 ? π§=0 π§=1 π§=2 100 115 130 IQ (π₯) You may be wondering why we have to look up the values in a table, rather than being able to calculate it directly. The reason is that calculating the area under the graph involves integrating (see C2), but the probability function for the normal distribution (which you won’t see here) cannot be integrated! Use of the z-table Suppose we’ve already worked out the z value, i.e. the number of standard deviations above or below the mean. 1 2 π§ π§ = −0.3 π§ = −0.3 π§ π π > −0.3 = π π < 0.3 = 0.6179? π π < −0.3 = 1 − π π < 0.3 = 0.3821? 3 4 This is clear by symmetry. π§=1 π π >1 =1−π π <1 ? = 0.1587 π§ π§ = −2 π π < −2 = 1 − π π < 2 = 0.0228? π§ Bro Tip: We can only use the z-table when: a) The z value is positive (i.e. we’re on the right half of the graph) b) We’re finding the probability to the left of this z value. Bro Tip: Either changing the sign of changing the direction of the inequality does “1 –”. If we do both, they cancel out. Test Your Understanding ? = π. ππππ π π > 0.70 = π − π· π < π. ππ ? = π. ππππ π π < −1.25 = π − π· π < π. ππ π π > −2.42 = π· π < π. ππ =? π. ππππ Exercise 1 Determine the following probabilities, ensuring you show how you have manipulated your probabilities (as per the previous examples). 1 2 3 4 5 6 7 8 9 10 π π π π π π π π π π π π π π π π π π π π < 1 = π. ππππ ? > 2.1 = 1 − π π < 2.1 ? = π. ππππ > −0.3 = π π < 0.3 =?π. ππππ < 1.5 = π. ππππ ? < −0.7 = 1 − π π < 0.7 = π.?ππππ < −1.7 = π π > 1.7 = 1 − π?π < 1.7 = π. ππππ > 0 = π. π ? < −0.4 = π π > 0.4 = 1 − π?π < 0.4 = π. ππππ > −2.4 = π π < 2.4 = π. ππππ ? > 3 = 1 − π π < 3 = π. ππππ ? ‘Standardising’ We have seen that in order to look up a value in the π§ table, we needed to first convert our IQ into a π§ value. We call this ‘standardising’ our variable, because we’re turning our normally distributed variable π into another one π where the mean is 0 and the standard deviation is 1. IQ (X) world Z world (recall that π = 100 and π = 15) π=3 π = 145 ? By thinking of what calculation you did, can you therefore come up with a formula for π in terms of π, π and π? !π = π−π ? π Bro Side Note: π~π 0,12 , but why? Well consider a z value of 3 for example. We understand that to mean 3 standard deviations above the mean. But if π = 0 and π = 1, the 3 is 3 lots of 1 above 0! Example The heights in a population are normally distributed with mean 160cm and standard deviation 10cm. Find the probability of a randomly chosen person having a height less than 180cm. Here’s how they’d expect you to lay out your working in an exam: π~π 160, 102 ? π π < 180 180 − 160 =π π< ? 10 = π π <? 2 = 0.9772 ? No marks attached with this, but good practice! A statement of the problem. M1 for “attempt to standardise” Look up in z-table Test Your Understanding Edexcel S1 May 2012 π~π 162,7.52 π π > 150 150 − 162 =π π> = π π > −1.6 7.5 = π π < 1.6 = 0.9452 ? Edexcel S1 May 2013 (R) π~π 150,102 π π < 145 145 − 150 =π π< = π π < −0.5 10 = 1 − π π < 0.5 = 0.3085 ? Probabilities for Ranges Again, let π represent the IQ of a randomly chosen person, where π~π 100,152 Thinking about the graph of the normal distribution, find: π· ππ < πΏ < πππ This easiest way is to find π π < 112 and ‘cut out’ the area corresponding to π π < 96 : π 96 < π < 112 = π π < 112 − π π < 96 112 − 100 96 − 100 =π π< −π π < 15 15 = π π < 0.8 − π π < −0.266 = π π < 0.8 − 1 − π π < 0.266 = 0.7881 − 1 − 0.6064 = 0.3945 ? z=0 96 100 IQ (π₯) 112 ! Probabilities of ranges: π· π<πΏ<π =π· πΏ<π −π· πΏ<π Test Your Understanding Let X represent the IQ of a randomly chosen person, where π~π 100,152 Use your Z-table to find: 1 P(100 < X < 107.5) = P(Z < 0.5) – 0.5 = 0.1915 ? 2 P(123 < X < 151) = P(1.53 < Z < 3.4) = P(Z < 3.4) ? – P(Z < 1.53) = 0.0627 3 P(70 < X < 90) = P(-2 < Z < -0.67) = (1 – 0.7486) ? – (1 – 0.9772) = 0.2286 Exercise 2 1 (On provided sheet) [Jan 2013 Q4a] The length of time, L hours, that a phone will work before it needs charging is normally distributed with a mean of 100 hours and a standard deviation of 15 hours. Find P(L > 127). = 0.0359 5 ? 2 [Jan 2012 Q7a] A manufacturer fills jars with coffee. The weight of coffee, W grams, in a jar can be modelled by a normal distribution with mean 232 grams and standard deviation 5 grams. Find P(W < 224). = 0.0548 ? 6 ? 3 [May 2011 Q4a] Past records show that the times, in seconds, taken to run 100 m by children at a school can be modelled by a normal distribution with a mean of 16.12 and a standard deviation of 1.60. A child from the school is selected at random. Find the probability that this child runs 100 m in less than 15 s. = 0.2420 [Jan 2011 Q8a] The weight, X grams, of soup put in a tin by machine A is normally distributed with a mean of 160 g and a standard deviation of 5 g. A tin is selected at random. Find the probability that this tin contains more than 168 g. = 0.0548 ? [May 2009 Q8a,b] The lifetimes of bulbs used in a lamp are normally distributed. A company X sells bulbs with a mean lifetime of 850 hours and a standard deviation of 50 hours. (a) Find the probability of a bulb, from company X, having a lifetime of less than 830 hours. = 0.3446 (b) In a box of 500 bulbs, from company X, find the expected number having a lifetime of less than 830 hours. = 172.3 ? ? ? 4 [May 2010 Q7a] The distances travelled to work, D km, by the employees at a large company are normally distributed with D οΎ N( 30, 82 ). Find the probability that a randomly selected employee has a journey to work of more than 20 km. = 0.8944 7 [Jan 2009 Q6a] The random variable X has a normal distribution with mean 30 and standard deviation 5.Find P(X < 39). = 0.9641 ? Exercise 2 (On provided sheet) 8 [May 2008 Q7a] A packing plant fills bags with cement. The weight X kg of a bag of cement can be modelled by a normal distribution with mean 50 kg and standard deviation 2 kg. Find P(X > 53). = 0.0668 ? 9 The heights of raccoons in a Canadian town are normally distributed with mean 30cm and standard deviation 5cm. Determine the probability of a randomly chose raccoon having a height between: a) 28cm and 31cm π· ππ < πΏ < ππ = π· −π. π < π < π. π = π. ππππ − π − π. ππππ = π. ππππ b) 33cm and 37cm π· ππ < πΏ < ππ = π· π. π < π < π. π = π. ππππ − π. ππππ = π. ππππ c) 25cm and 28cm π· ππ < πΏ < ππ = π· −π < π < −π. π = (π − π. ππππ) − π − π. ππππ = π. ππππ ? ? ? The reverse: Finding the z-value for a probability Sometimes we’re given the probability, and need to find the π§ value, so that we can determine a missing value or the standard deviation/mean. Just use the z-table backwards! π π π π π π π π π π π π π π π π π π π π π π π π π π <π§ <π§ <π§ <π§ >π§ <π§ <π§ <π§ <π§ <π§ >π§ >π§ >π§ = 0.87 = 0.92 = 0.7 = 0.85 = 0.23 = 0.1 = 0.95 = 0.7 = 0.46 = 0.01 = 0.86 = 0.975 = 0.43 → π§ = 1.13 ? → π§ = 1.41 ? → π§ = 0.5244 ? → π§ = 1.0364 ? → π π < π§ = 0.77 ? → π π < −π§ = 0.9 ? → → → → → → → π§ = 1.6449 ? π§ = 0.5244 ? π π < −π§ = 0.54? π π < −π§ = 0.99 ? π π < −π§ = 0.86 ? π π < −π§ = 0.975? π π < π§ = 0.57 ? For nice ‘round’ probabilities, we have to look in the second z-table. You’ll lose a mark otherwise. π§ = 0.74 π§ = −1.2816 Bro Tip: Remember that either flipping the inequality, or changing the sign of π§ will cause your probability to become 1 minus it. → → → → → π§ = −0.10 π§ = −2.3263 π§ = −1.08 π§ = −1.9600 π§ = 0.18 Dealing with two-ended inequalities Bro Tip: The key is to use a diagram (or otherwise) to convert to a single-ended inequality. Sometimes the range is two ended. These can be one of two possible forms in an exam: Find the value of π such that π 0 < π < π = 0.3 Find the value of π such that π −π < π < π = 0.7 ? ? Suitable Diagram 0.3 0.7 Diagram Suitable 0.15 π −π 0.15 π This can be simplified to: π· π< ? π = π.?π This can be simplified to: π· π< ? π = π.?ππ Thus: Thus: π = π. ππππ ? π = π. ππππ ? Test Your Understanding 1 Find the value of π such that π −π < π < π = 0.9 π· π < π = π. ππ ? π = π. ππππ 2 Find the value of π such that π π < π < 0 = 0.4 π· π < π = π. π π· π < −π = π. π −π = π. ππππ ? π = −π. ππππ We’ll come back to this later… Where we are so far “IQ has the distribution π~π 100,152 An IQ of 130 corresponds with what z-value?” πΏ−π πππ−πππ π = π = ?ππ = π οΌ Understand what a Z-value means and the formula to calculate it. οΌ Calculate a probability of being above or below a particular value. οΌ “Find the probability that I have an IQ above 115.” π· πΏ > πππ = π· π > π =? π−π· π<π = π. ππππ Calculate a z-value corresponding to a probability. π π π π π π π π <π§ >π§ <π§ >π§ = 0.58 = 0.22 = 0.15 = 0.95 π = π. ππ ? π = π. ππ ? π = −π. ππππ ? π = −π. ππππ ? COMING NOW! Calculate a value corresponding with a probability. “Find the IQ corresponding with the bottom 30% of the population.” COMING SOON! Calculate a missing value of π and/or π. Solve more complex problems (e.g. involving conditional probabilities) Retrieving the original value Again, let X represent the IQ of a randomly chosen person, where π~π 100,152 What IQ corresponds to the bottom 78% of the population? State the problem in probabilistic terms. π π < π₯? = 0.78 π π < π§? = 0.78 ‘Standardise’. 0.78 z = 0.77 Bro Tip: Draw a diagram for these types of questions if it helps. π§ =? 0.77 Identify z value (as we did in the previous lesson). Use z formula to find π₯ π₯ − 100 = 0.77 15 ? π₯ = 100 + 0.77 × 15 = 111.55 Retrieving the original value Again, let X represent the IQ of a randomly chosen person, where π~π 100,152 What IQ corresponds to the bottom 90% of the population? π π<? π₯ = 0.9 State the problem in probabilistic terms. ‘Standardise’ π π < π§? = 0.9 0.9 Identify z value z = 1.2816 z = 1.2816 ? Use z formula to find π₯ π₯ − 100 = 1.2816 15 ? π₯ = 119.224 Retrieving the original value Again, let X represent the IQ of a randomly chosen person, where π~π 100,152 What IQ corresponds to the bottom 30% of the population? State the problem in probabilistic terms. π π < π₯? = 0.3 π π < π§? = 0.3 π π < −π§ = 0.7 −π§ = 0.5244 ? π§ = −0.5244 ‘Standardise’ Identify z value Use z formula to find π₯ π₯ − 100 ? 15 π₯ = 92.143 −0.5244 = Retrieving the original value Again, let X represent the IQ of a randomly chosen person, where π~π 100,152 What IQ does 80% of the population have a value more than? State the problem in probabilistic terms. π π > π₯? = 0.8 ‘Standardise’ π π > π§? = 0.8 π π < −π§ = 0.8 −π§ = 0.8416 ? π§ = −0.8416 Identify z value Convert back into an IQ. π₯ − 100 ? 15 π₯ = 87.376 −0.8416 = Test Your Understanding [May 2008 Q7b] A packing plant fills bags with cement. The weight X kg of a bag of cement can be modelled by a normal distribution with mean 50 kg and standard deviation 2 kg. Find the weight that is exceeded by 99% of the bags. Remember: (5) 1. 2. ? 3. 4. State your problem in probabilistic terms. Standardise. Manipulate if necessary so that your probability is above 0.5 and you’re finding π π < π ππππ‘βπππ . Use your z-table backwards to find the zvalue. π₯−π Use π§ = to find your value of π₯. π [May 2011 Q4b] Past records show that the times, in seconds, taken to run 100 m by children at a school can be modelled by a normal distribution with a mean of 16.12 and a standard deviation of 1.60. On sports day the school awards certificates to the fastest 30% of the children in the 100 m race. Estimate, to 2 decimal places, the slowest time taken to run 100 m for which a child will be awarded a certificate. (4) ? Additional Practice (Outside of Class) On the planet Frostopolis, the mean height of a Frongal is 1.57m and the standard deviation 0.2m. Determine: a) The height for which 65% of Frongals have a height less than. π π < π₯ = 0.65 π₯ − 1.57 π π< = 0.65 0.2 π₯ − 1.57 = 0.39 0.2 π₯ = 1.648 π b) The height for which 40% of Frongals have a height more than. ? c) The height for which 23% of Frongals have a height less than. π π < π₯ = 0.23 π π < π§ = 0.23 π π < −π§ = 0.77 π§ = −0.74 π₯ − 1.57 = −0.74 0.2 π₯ = 1.422 π ? π π > π₯ = 0.4 π π > π§ = 0.4 π π < π§ = 0.6 π₯ − 1.57 = 0.2533 0.2 π₯ = 1.62 π ? Remember: 1. State your problem in probabilistic terms. 2. Standardise. Manipulate if necessary so that your probability is above 0.5 and you’re finding π π < π§ . 3. Use your z-table backwards to find the z value. π₯−π 4. Use π§ = to find your value of π₯. π Exercise 3a (On provided sheet) 1 [Jan 09 Q6b-d] The random variable X has a normal distribution with mean 30 and standard deviation 5. (b) Find the value of d such that P(X < d) = 0.1151. (4) π = ππ ? (c) Find the value of e such that P(X > e) = 0.1151. (2) π = ππ ? (d) Find P(d < X < e). (2) π. ππππ ? 2 [Jan 2013 Q4b] The length of time, L hours, that a phone will work before it needs charging is normally distributed with a mean of 100 hours and a standard deviation of 15 hours. Find the value of d such that P(L < d) = 0.10. (3) π = ππ. πππ ? 3 [Jan 2011 Q8b] The weight, X grams, of soup put in a tin by machine A is normally distributed with a mean of 160 g and a standard deviation of 5 g. The weight stated on the tin is w grams. Find w such that P(X < w) = 0.01. (3) π = πππ. ? ππ 2 4 [Ex9C Q6] Given π~π 30, 5 , find π such that π π > π = 0.30 π = ππ. π ? [Ex9C Q7] Given π~π 15, 32 , 5 find π such that π π > π = 0.15 π = ππ. π ? Exercise 3a (On provided sheet) 6 [Jan 2007 Q7b] The measure of intelligence, 8 [Solomon Paper E Q3c] The random variable π is normally distributed with a IQ, of a group of students is assumed to be mean of 42 and a variance of 18. Find Normally distributed with mean 100 and the value of π₯ such that standard deviation 15. π π ≤ π₯ = 0.95 The probability that a randomly selected π = ππ. π student as an IQ of at least 100 + k is 0.2090. ? Find, to the nearest integer, the value of k. (6) 2 π =?ππ 9 [Ex9C Q9] Given π~π 100,15 . Find the value of π and π such that: [June 2005 Q6b,c] A scientist found that the 7 a) π π < π = 0.40 time taken, M minutes, to carry out an experiment can be modelled by a normal random variable with mean 155 minutes and standard deviation 3.5 minutes. (b)P(150 ≤ M ≤ 157), (4) π. ππππ ? (c) the value of m, to 1 decimal place, such that P(M ≤ m) = 0.30. (4) π = πππ. ? π ? π π = ππ. b) π π > π = 0.6915 ?π π = ππ. c) π π < π < π = π. ππππ ? Harder reverse probability questions Recap: Find the value π such that π −π < π < π = 0.7 π· π < π = π. ππ ? π = π. ππππ Q: If π~π 30,52 , find the value of π such that π 30 < π < π = 0.2 Observe that the 30 at the bottom end of the inequality is the mean. Thus: π· πΏ < π = π. π π − ππ ? = π. ππππ π π = ππ. πππ Quartiles and Percentiles Using IQ: π~π 100, 152 Find: The Lower Quartile π1 ?π = π.?ππ π· πΏ<πΈ π· π < π = π. ππ π· π < −π = π. ππ πΈπ − πππ −π. ππ = ? ππ πΈπ = ππ. ππ The 70th percentile π70 i.e. 25% of people have an IQ less than the lower quartile. The Upper Quartile π3 π· πΏ < πΈπ = π. ππ π· π < π = π. ππ πΈπ − πππ π. ππ = ? ππ πΈπ = πππ. ππ π· πΏ < π·ππ = π. π π·ππ − πππ π. ππππ = ? ππ πΈπ = πππ. πππ Test Your Understanding Edexcel S1 May 2012 ? Edexcel S1 May 2010 a) b) c) P(D > 20) = P(Z > -1.25) = P(Z < 1.25) = 0.8944 P(Z < z) = 0.75 -> z = 0.67 Q3 = 30 + (0.67 x 8) = 35.36 Q1 = 30 – (0.67 x 8) = 24.64 ? Exercise 3b (On provided sheet) Finding missing π and π The random variable π~π π, 32 Given that P(X > 20) = 0.20, find the value of π. π(π < 20) = 0.8 π π < π§ = 0.8 (Standardising) π§ = 0.8416 ? 20 − π = 0.8416 3 π = 17.5 The random variable π~π 50, π 2 Given that P(X < 46) = 0.2119, find the value of π. π π < 46 = 0.2119 π π < π§ = 0.2119 π π < −π§ = 0.7881 π§ = −0.8 46 − 50 ?= −0.8 π π=5 If your standard deviation is negative, you know you’ve done something wrong! Finding missing π and π The random variable π~π π, π 2 Given that P(X > 35) = 0.025 and P(X < 15) = 0.1469, find the value of π and the value of π First deal with P(X > 35) = 0.025 Next deal with P(X < 15) = 0.1469 P(X < 35) = 1 – 0.025 = 0.975 If P(Z < z) = 0.975, then z = 1.96. π π < 15 = 0.1469 π π < π§ = 0.1469 π π < −π§ = 0.8531 π§ = −1.05 35 − π? = 1.96 π π + 1.96π = 35 ? 15 − π = −1.05 π π − 1.05π = 15 We now have two simultaneous equations. Solving gives: π = 22.0 π =?6.64 Test your understanding For the weights of a population of squirrels, which are normally distributed, Q1 = 0.55kg and Q3 = 0.73kg. Find the standard deviation of their weights. ο³ = 0.114 ο» ο³ = 0.124 ο» ο³ =οΌ 0.134 ο³ = 0.144 ο» Due to symmetry, ο = (0.55 + 0.73)/2 = 0.64kg If P(Z < z) = 0.75, then z = 0.67. 0.64 + 0.67ο³ = 0.73 ο³ = 0.134 Only 10% of maths teachers live more than 80 years. Triple that number live less than 75 years. Given that life expectancy of maths teachers is normally distributed, calculate the standard deviation and mean life expectancy. π· πΏ < ππ = π. π ο = 76.15 ο» ο = 76.25 ο» ο = 76.35 ο» ο =οΌ 76.45 ο³ =οΌ 2.77 ο³ =ο» 2.78 ο³ =ο»79 ο³ =ο» 2.80 π· π < π = π. π π = π. ππππ ππ − π = π. ππππ π π + π. πππππ = ππ Similarly ο – 0.5244ο³ = 75 Exam Questions Edexcel S1 May 2013 (R) ? Edexcel S1 Jan 2011 ? Edexcel S1 Jan 2002 a) z-value for 0.975 is 1.96. By symmetry, 235 is 1.96 standard deviations below mean. So π − 1.96π = 235. The result follows. b) P(Z < z) = 0.85. So z = 1.04 π + 1.04π = 286 c) Solving, π = 268.32, π = 17 d) If 0.683 in the middle, (0.683/2)+0.5=0.8415 prob below value above mean. Thus z = 1. So values are 154.8 – 2.22 and 154.8 + 2.22 ? Summary A B A normal distribution is good for modelling data which: tails off symmetrically about some mean/ has a bell-curve like ? distribution. A z-value is: The number of standard deviations above ? the mean. C If a random variable X is normally distributed with mean 50 and standard deviation 2, we would write: πΏ~π΅ ?ππ, ππ D A z-table is: A cumulative distribution function for a normal distribution with ? mean 0 and standard deviation 1. P(IQ < 115) = 0.8413 ? E We need to use the second z-table whenever: we’re looking up the z value for certain ‘round’ probabilities. ? F P(a < X < b) = P(X < b) –?P(X < a) G We can treat quartiles and percentiles as probabilities. For IQ, what is the 85th percentile? 100 + (1.04 x?15) = 115.6 H We can form simultaneous equations to find the mean and standard deviation, given known values with their probabilities. Mixed Questions Edexcel S1 June 2001