CHAPTER 8.

Interval Estimation

- A sample statistic is used to make statistical inferences regarding the unknown value of
the population parameter.
- When you make an estimate, there is always uncertainty around that estimate because the
number is based on a sample for the population you are studying.

-------------------------------------------------------------------------------------------------------------------

Confidence Interval For the Population Mean When σ Is Known

When a statistic is used to estimate a parameter, it is referred to as a point estimator.

- Sample mean, x̅ , estimator of population mean, 𝑢

- Sample proportion, 𝑝̅, estimator of the population proportion, p

-------------------------------------------------------------------------------------------------------------------

Confidence Intervals

- The range of values that you expect your estimate to fall between some percent of the
time, IF you re-run your experiment OR re-samples the population in the same way.
- It is the percentage of times you expect to reproduce an estimate between the upper and
lower bounds of the confidence interval and is set by the alpha value.
- Eg 95% confidence level = 95 out of 100 times the estimate will fall between the
upper and lower values specified by the confidence interval.
- It is essential that the sampling distributions of 𝑋 ̅ and 𝑃 ̅ follow a normal distribution.
- Mean: underlying population is normal or 𝑛≥30
- Proportion: 𝑛𝑝 ≥ 5 and 𝑛 (1−𝑝) ≥ 5
- Confidence intervals have a margin of error that accounts for:
- The standard error of the estimator
- The desired confidence level of the interval
 - The confidence interval for the population mean and population proportion is constructed
as: point estimate ± margin of error.

-------------------------------------------------------------------------------------------------------------------

Confidence Interval ( 1- 𝛼 )

- 1- 𝛼 is the significance level.

- Eg construct a confidence interval with 90% confidence.
- 1- 𝛼 = 1- 0.90 = 0.10, therefore, alpha = 0.10

-------------------------------------------------------------------------------------------------------------------

Z(α / 2 )

- The 𝑧 value with the probability of 𝛼/2 in the upper tail of the standard normal
probability distribution.
- The area under the Z curve to the right of the value

-------------------------------------------------------------------------------------------------------------------

Confidence level, alpha and Z- score.

- 95% confidence, 𝛼 = 0.05, 𝛼 ∕ 2 = 0.025, Z0.025 = 1.96

- 90% confidence, 𝛼 = 0.10, 𝛼 ∕ 2 = 0.05, Z0.05 = 1.645
- 99% confidence, 𝛼 = 0.01, 𝛼 ∕ 2 = 0.005, Z0.005 = 2.576

-------------------------------------------------------------------------------------------------------------------

Margin of Error

- ( Z(α / 2 ) ) (σ / √n )

-------------------------------------------------------------------------------------------------------------------
Sample Size Effects on CI Width

- CI width decreases (narrower): as sample size increases, standard error decreases AND /
OR margin of error decreases
- CI width increases: as sample size decreases, standard error increases AND / OR margin
of error increases

-------------------------------------------------------------------------------------------------------------------

How confidence level (CL) affects CI size/width

- Increased CL = increased CI width

- Lower CL = decreased CI width

-------------------------------------------------------------------------------------------------------------------

How Standard Deviation affects CI Width

- Increased standard deviation = increased CI width

-------------------------------------------------------------------------------------------------------------------

-------------------------------------------------------------------------------------------------------------------

CHAPTER 8.2

Confidence Interval for the Population Mean When σ Is Unknown

- Margin of error depends on the standard error of the estimator.

- When 𝜎 is unknown, use the sample standard deviation 𝑠
- Estimate 𝜎 ∕ √𝑛 with 𝑠 ∕ √𝑛

-------------------------------------------------------------------------------------------------------------------

Standard Deviation is Unknown : T-Test

- Find tdf
- tdf = (x̅ − 𝜇) / ( 𝑠 ∕ √𝑛 )
-------------------------------------------------------------------------------------------------------------------

t distribution

- Similar to Z distribution
- Bell-shaped and symmetric around zero
- Slightly broader tails than Z distribution
- Identified by the degrees of freedom, df
- Fewer df = broader tails
- As df increases, t becomes more like Z

-------------------------------------------------------------------------------------------------------------------

Degrees of Freedom

- Referred to as 𝑡𝑑𝑓 distribution

- df = (𝑛−1)
- ie Number of independent pieces of information that can be freely chosen
- Lose one degree of freedom for computing the sample mean

-------------------------------------------------------------------------------------------------------------------

T-Table

- t table is not as comprehensive as z table

- Only lists probabilities for a limited number of values in the upper tail

-------------------------------------------------------------------------------------------------------------------

Confidence Interval

- CI = 𝑥̅ ± ( 𝑡(𝛼 ∕ 2, 𝑑𝑓) ) (𝑠 / √𝑛 )
- Only valid when X follows a normal distribution.
- Uncertainty increases when using the sample standard, making the confidence interval
wider.
- This is denoted by the wider tail of the t α / 2 ,df distribution.
-------------------------------------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------------------------------------

CHAPTER 8.3

Confidence Interval for the Population Proportion

- p represents the proportion of successes in the population.

- p̅ is approximately normally distributed when 𝑛𝑝 ≥ 5 and 𝑛 ( 1−𝑝 ) ≥ 5.
- CI = p̅ ± ( Z(𝛼∕2) ) ( √( 𝑝̅ (1− 𝑝̅ ) / 𝑛 )

-------------------------------------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------------------------------------

CHAPTER 9.1

Introduction to Hypothesis Testing

- Hypothesis testing is a statistical inference.

- Used to determine whether a statement about the population parameter should or
should not be rejected.
- Used to resolve conflicts between two competing hypothesis on a particular
population of interest.

-------------------------------------------------------------------------------------------------------------------

Null and Alternative Hypothesis

- Null hypothesis 𝐻0
- Status quo or “business as usual”
- Whatever you are trying to DISPROVE should be the NULL hypothesis.
- Specified with =, ≤ or ≥
- Alternative hypothesis 𝐻A
- Contests status quo
- Whatever we wish to establish, something new
- What you would like to support will be your alternative hypothesis.
 - Specified with the opposite of what is in the null: ≠, > or <

-------------------------------------------------------------------------------------------------------------------

Formulating a Hypothesis Test

- 3 general steps:
- Identify the population parameter of interest.
- Determine whether it is a one or two tailed test.
- Include some form of equal sign in the null hypothesis and use the alternative
hypothesis to establish a claim.

-------------------------------------------------------------------------------------------------------------------

Type I and Type II Errors

- Type I Error – rejecting the null when we should not reject the null (false positive)
- Type II Error – not rejecting the null when we should reject the null (false negative)

-------------------------------------------------------------------------------------------------------------------

- Common practice is to control or minimize Type I error rate.

- Don’t reject the null unless confident that rejection is the correct decision.
- ie don’t reject the null unless the sample stat is relatively far from the
hypothesized parameter.

-------------------------------------------------------------------------------------------------------------------

Level of significance α

- Determined by defining a max allowable Type I error rate.

- Common Levels: 0.01, 0.05 and 0.10

-------------------------------------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------------------------------------

CHAPTER 9.2
Hypothesis Test for the Population Mean when 𝜎 is Known

- 𝜎 is rarely known.
- Instances where 𝜎 is stable and can be determined from prior experience
- In these cases, treat 𝜎 as known.
- Assume the null hypothesis is true.
- And determine if the sample evidence contradicts this assumption.

-------------------------------------------------------------------------------------------------------------------

Sample mean (x̅ )

- A test about 𝜇 is based on x̅ (sample mean).

- E(x̅) = 𝜇 (expected value)
- se(x̅) = 𝜎 ∕ √𝑛 (standard error)
- Sampling distribution of 𝑋 ̅ is (approximately) normal (underlying population is normal
or 𝑛≥30).

A large difference between 𝑥̅ and 𝜇 does not mean the null hypothesis is false, it could be
explained by chance.

-------------------------------------------------------------------------------------------------------------------

Test Statistic

- Accounts for the variability

- 𝜇 = 𝜇0
- Z = ( 𝑥̅ − 𝜇0 ) / ( 𝜎 ∕ √𝑛 )
- Used to find the critical value and p-value.

-------------------------------------------------------------------------------------------------------------------

Lower Tail Test

- Ho: µ ≥ 𝜇0
- Ha: µ < 𝜇0
Upper Tail Test

- Ho: µ ≤ 𝜇0
- Ha: µ > 𝜇0

Two Tail Test

- Ho: µ = 𝜇0
- Ha: µ ≠ 𝜇0

-------------------------------------------------------------------------------------------------------------------

Summary of Steps

- Determine and describe the appropriate sampling distribution

- Find the critical values based on α and establish rejection rule.
- Calculate the test statistic from sample.
- Compare test statistic and critical value.
- Use rejection rule to either reject or do not reject the null hypothesis.
- Find your test stat then find the critical value to decide if you reject or do not
reject.
- Find your test stat then find the p-value to decide if you reject or do not reject.
- Interpret your results

Critical value is the corresponding Z value

-------------------------------------------------------------------------------------------------------------------

P-Value to Determine Rejection or Do Not Reject the Null

- A probability that provides a measure of evidence against the null hypothesis provided by
the sample.
- Smaller p-values = more evidence against the null hypothesis.
- 𝛼 = the allowed probability of making a Type I error
- Choose 𝛼 before implementing a test
 - Typically, 𝛼 = 0.01, 0.05, or 0.10
- p-value is the observed probability of making a Type I error.
- If p-value < 𝛼 reject the null hypothesis
- If p-value ≥ 𝛼 do not reject the null hypothesis

p-value and critical value do more or less the same thing, but p-value uses the test statistic while
critical value uses the Z – score

-------------------------------------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------------------------------------

CHAPTER 9.3

Hypothesis Test for the Population Mean when 𝜎 is Unknown

- In most applications 𝜎 not known.

- The sampling distribution of x̅ is (approximately) normal (underlying population is
normal or 𝑛 ≥ 30).

-------------------------------------------------------------------------------------------------------------------

Test Statistic

- Let 𝜇₀ be the hypothesized mean.

- Test statistic = 𝑡𝑑𝑓 = (𝑥̅ − 𝜇₀) / ( 𝑠 ∕ √ 𝑛 ) and 𝑑𝑓 = 𝑛−1.

What is df?

- df = n – 1
- df ≡ degrees of freedom → the number of independent pieces of information used to
calculate a statistic.
- Degrees of freedom used to justify the results to accept or reject the null.
- Determines the shape of the t-distribution which is used to calculate the probability
of observing a certain sample result: Reject or Do Not Reject the Null
- We need the degrees of freedom to find the p-value in the T-Table
-------------------------------------------------------------------------------------------------------------------

T-table

- We find our test stat and use the absolute value.

- Meaning we will always make our test stat positive
- Your test stat most likely will not show up in the table
- You will need to find the closest or which values it will be between

-------------------------------------------------------------------------------------------------------------------

-------------------------------------------------------------------------------------------------------------------

CHAPTER 9.4

Hypothesis Test for the Population Mean Proportion

- The population proportion 𝑝 represents the proportion of observations with a particular

attribute.
- Estimate population proportion using the sample proportion, p̅
- p̅ has an approximate normal distribution when 𝑛𝑝 ≥ 5 and 𝑛 ( 1− 𝑝 ) ≥ 5

-------------------------------------------------------------------------------------------------------------------

Test Statistic

- Let 𝑝₀ be the hypothesized value of the population proportion in the null hypothesis.
- The test statistic: Z = (𝑝̅ − 𝑝₀) / √( 𝑝₀ ( 1 − 𝑝₀ ) ∕ 𝑛 )

-------------------------------------------------------------------------------------------------------------------

- Forms of Null and Alternative Hypotheses

Ho: p = po versus Ha: p ≠ po
- Ho: p ≤ po versus Ha: p > po
- Ho: p ≥ po versus Ha: p < po
j