The Normal Distribution & z-scores

Transcription

1 & z-scores

2 Distributions: Who needs them? Why are we interested in distributions? Important link between distributions and probabilities of events If we know the distribution of a set of events, then we know something about the probability that one of those events is likely to occur

3 Distributions and Probabilities Whose course would you rather take? Professor A Professor B C/C+ 30% D 8% F 5% A 19% B/B+ 38% D 4% C/C+ 22% F 7% B/B+ 35% A 32%

4 Distributions and Probabilities Whose course would you rather take? Professor A Professor B D 4% F 5% D 8% C 30% B 38% A 19% F 7% C 22% B 35% A 32%

5 A Note on Continuous Versus Discrete Distributions Y-axis represents density Probability density function Probability of any exact value is zero Y-axis represents probability Probability mass function Probability of exact values can be positive

6 The Normal Distribution Okay, distributions can be useful. But why should we care about this particular distribution? Most prominent distribution in statistics Normally distributed data Computational convenience Normality of sample means

7 Normally Distributed Data Many of the dependent variables that we deal with are approximately normally distributed in the population Originally used to describe errors in astronomical measurements Examples of normally distributed data Height Weight Normalized test scores Speed of cars on Route 18 Anything averaged across a large number of observations

8 Normally Distributed Data

9 Computational Convenience Many statistics related to the normal distribution, along with their sampling distributions are analytically tractable If we can assume that a variable is at least approximately normally distributed, then we can use standard techniques (i.e., those that make up most of this book) to make inferences about values of that variable

10 Normality of Sample Means Regardless of the distribution of the underlying variable (with limited exceptions), the distribution of sample means approaches normality as the sample size n grows. We will discuss this in greater detail later in the course when we cover the central limit theorem. Generally, this means that the sampling distribution of the mean can be approximated by a normal distribution

11 The Standard Normal Distribution The Normal Distribution σ 2.1% 13.6% 34.1% 34.1% 13.6% 2.1%

12 Standardizing Normal Variables By itself, a raw score or x value provides very little information about how that particular score compares with other values in the distribution. A score of x = 53, for example, may be a relatively low score, or an average score, or an extremely high score depending on the mean and standard deviation for the distribution from which the score was obtained. If the raw score is transformed into a z-score, however, the value of the z-score tells you where the score is located relative to all the other scores in the distribution.

13 Standardizing Normal Variables To transform an x value into a z-score: z x µ = σ To transform a z-score into an x value: x = µ + zσ

14 Standardizing Normal Variables In addition to knowing the basic definition of a z-score and the formula for a z-score, it is useful to be able to visualize z- scores as locations in a distribution. Remember, z = 0 is in the center (at the mean), and the extreme tails correspond to z-scores of approximately 2.00 on the left and on the right. 95.4% of the distribution is contained between z = -2 and z = % of the distribution is contained between z = -3 and z = 3

15 Standardizing Normal Distributions The Normal Distribution z = ( x µ ) : ( x µ ) : σ

16 Standard Normal Tables (z-tables) Another important advantage of standardizing distributions is that it allows us to compute and use a single probability table for all normal distributions The Normal Distribution

17 Standard Normal Tables (z-tables) Upper-Tail Probabilities The Normal Distribution z

18 Standard Normal Tables: Meaning The Normal Distribution

19 Standard Normal Tables: Meaning The Normal Distribution

20 Standard Normal Tables: Meaning The Normal Distribution

21 Standard Normal Tables: Meaning The Normal Distribution

22 Standard Normal Tables: Meaning The Normal Distribution

23 Using z-tables Area under curve sums to 1 The normal distribution is symmetrical E.g., P(Z < -z) = P(Z > +z) (non-overlapping) areas sum E.g., P(0.0 < Z < 1.0) = P(Z > 0.0) P(Z > 1.0) Note: In this slide and in the rest of this lecture, I m using Z (capital) to indicate a randomly selected value of a standard normal variable, and z (lower-case) to indicate a particular value of that variable.

24 What is the probability that a randomly selected woman is taller than 5 6 (66 inches)?

25 Using z-tables to Compute Interval Probabilities Step 1: standardize value (i.e., compute z-score) 66 µ z(66) = = σ Step 2: Look up upper-tail area P( Z > z(66) ) =

26 Standard Normal Tables (z-tables) Upper-Tail Probabilities The Normal Distribution z

27 What is the probability that a randomly selected woman is between 64 and 68 inches tall?

28 Using z-tables to Compute Interval Probabilities Step 1: standardize values (i.e., compute z-scores) 64 µ z(64) = = σ 68 µ z(68) = = σ Step 2: Look up upper-tail area ( z(64) ) ( z(68) ) P Z P Z > = > = Step 3: Compute difference ( ) ( (64)) ( (68)) P < Z < = P Z > z P Z > z =

29 Standard Normal Tables (z-tables) Upper-Tail Probabilities The Normal Distribution z

30 What is the probability that a randomly selected woman is between 62 and 66 inches tall?

31 Standard Normal Tables (z-tables) Upper-Tail Probabilities The Normal Distribution z

32 Computing Probable Intervals 95% 2.5% 2.5%

33 Standard Normal Tables (z-tables) Upper-Tail Probabilities The Normal Distribution z

34 Using z-tables to Compute Probable Intervals Step 1: look up z-score corresponding to 2.5% (0.0250) ( > z) = P Z z z + = 1.96 = z = Step 2: transform z-scores to heights x x = z σ + µ = = = z σ + µ = = 58.61