Chapter 8 & (part) Chapter 12: Distribution of Sample

Means

We’ve been examining Z-scores & the probability of obtaining

individual scores within a normal distribution

But inferential statistics involve samples of more than 1

To transition into inferential statistics, it is important that we understand

how probability relates to sample means, not just individual scores

Inferential statistics: sample  infer something about population

Often not possible to measure everyone in a population

Samples are convenient representations of them

Chapters 8 & 12: Page 1

If you take multiple samples of the same size from a population, they are
likely to give different results

Samples vary!

Quite likely that a particular sample won’t reflect the population exactly

Discrepancy b/n sample & population = sampling error

The term “sampling error” does not mean a sampling mistake – rather it
indicates that means drawn from multiple samples taken from a
population will vary from each other due to random chance and
therefore may deviate from the population mean

Chapters 8 & 12: Page 2

What is a distribution of Sample Means (Sampling Distribution of
the Mean)?

A distribution of sample means ( X ); a “distribution of a statistic [in this

case
a sample mean] over repeated sampling from a specified population”

Based on all possible random samples of size n, from a population

Can inform us of the degree of sample-to-sample variability we should expect

due to chance

Chapters 8 & 12: Page 3

Suppose we have a population: 6 7 8 9 µ = 7.5

Let’s take all possible samples of size n = 2 from this population:

1. 6 6 X = 6.0 7. 7 8 X = 7.5 13. 9 6 X

= 7.5
2. 6 7 X = 6.5 8. 7 9 X = 8.0 14. 9 7 X
= 8.0
3. 6 8 X = 7.0 9. 8 6 X = 7.0 15. 9 8 X
= 8.5
4. 6 9 X = 7.5 10. 8 7 X = 7.5 16. 9 9 X
= 9.0
5. 7 6 X = 6.5 11. 8 8 X = 8.0
6. 7 7 X = 7.0 12. 8 9 X = 8.5

What do you notice?

Chapters 8 & 12: Page 4

 X is rarely exactly µ
But, most X are close to (or cluster around) µ
Extreme values of X are rare
You can determine the exact probability of obtaining a particular X
p( X < 7)? = 3 / 16

Important properties of the sampling distribution of means:

1. Mean
2. Standard Deviation
3. Shape

1. The Mean

The mean of the distribution of sample means is the mean of the population

Chapters 8 & 12: Page 5

The mean of the distribution of sample means is called expected value of
X

X is an unbiased estimate of µ : on average, the sample mean produces a

value that exactly matches the population mean

Chapters 8 & 12: Page 6

2. The Standard Deviation of the Distribution of Sample Means

σ X = Standard Error of the Mean

σ
σ X=
n

Variability of X around µ

Special type of standard deviation, type of “error”

Average amount by which X deviates from µ

Less error = better, more reliable estimate of population parameter

Chapters 8 & 12: Page 7

σ X influenced by two things:
(1) Sample size (n)

Larger n = smaller standard errors

Note: when n = 1  σ X = σ

σ as “starting point” for σ X ,

σ X gets smaller as n increases

(2) Variability in population (σ )

Larger σ = larger standard errors

Chapters 8 & 12: Page 8

 3. The Shape

Central Limit Theorem = Distribution of sample means will approach a

normal distribution as n approaches infinity

Very important!
True even when raw scores NOT normal!
What about sample size?

(1) If raw scores ARE normal, any n will do

(2) If raw scores are not normal but are symmetrically distributed, a small
n will usually suffice

(3) If the raw scores are severely skewed, n must be “sufficiently large”

For most distributions  n ≥ 30

Chapters 8 & 12: Page 9

 Why are Sampling Distributions Important?
• Tell us the probability of getting a particular X , given µ & σ

• Critical for inferential statistics!

• Allow us to estimate population parameters

• Allow us to determine if a sample mean differs from a known population mean

just because of chance

• Allow us to compare differences between sample means – due to chance or

to experimental treatment?

Sampling distribution is the most fundamental concept underlying all statistical

tests

Chapters 8 & 12: Page 10

Working with the Distribution of Sample Means

If we assume DSM is normal (again, we can do this if raw scores are normally
distributed or n is at least 30)
AND
If we know µ & σ

We can use the Normal Curve & Table E.10!

x −µ
z=
σx

σ
where: σ X =
n

Chapters 8 & 12: Page 11

Example: Suppose you take a sample of 25 high-school students, and measure
their IQ. Assuming that IQ is normally distributed with µ = 100 and σ = 15,
what is the probability that your sample’s IQ will be 105 or greater?
x −µ
Step 1: Convert to Z-score: z=
σx

15 15
σ X= = =3
25 5
105 −100
z= =1.67
3

Step 2: See Table E. 10

The probability of the sample having a mean of 105 or greater is: 0.0475
Example: Repeat the same problem in the previous example, but assume your
sample size is 64

Chapters 8 & 12: Page 12

 x −µ
Step 1: Convert to Z-score: z=
σx

15 15
σ X= = =1.875
64 8
105 −100
z= = 2.67
1.875

Step 2: See Table E. 10

The probability of the sample having a mean of 105 or greater is: 0.0038

Chapters 8 & 12: Page 13

Example: What X marks the point above which sample means are likely to
occur only 15% of the time, if n = 36?

σ
Step 1: Find σ X =
n

15 15
σ X= = = 2.5
36 6

Step 2: See Table E. 10 : Z = ?

Step 3: Solve for X : X = µ + Zσ X

X = 100 + (?)(2.5) = 102.6

Chapters 8 & 12: Page 14