If there is one prayer that you should pray/sing every day and every hour, it is the LORD's prayer (Our FATHER in Heaven prayer)
It is the most powerful prayer. A pure heart, a clean mind, and a clear conscience is necessary for it.
- Samuel Dominic Chukwuemeka

God is our refuge and our strength, our ever-present help in distress. - Psalm 46:1

The Joy of a Teacher is the Success of his Students. - Samuel Chukwuemeka

Hypothesis Tests

Welcome to Our Site

I greet you this day,

First: Review the Notes.
Second: View the Videos.
Third: Solve the questions/solved examples.
Fourth: Check your solutions with my thoroughly-explained solved examples.

I wrote the codes for the calculators using JavaScript, a client-side scripting language and AJAX, a JavaScript library. Please use the latest Internet browsers. The calculators should work.

Comments, ideas, areas of improvement, questions, and constructive criticisms are welcome. You may contact me. If you are my student, please do not contact me here. Contact me via the school's system.

Samuel Dominic Chukwuemeka (SamDom For Peace) B.Eng., A.A.T, M.Ed., M.S

Objectives

Students will:

(1.) Discuss the meaning of hypothesis testing.
(2.) Explain the meaning of null and alternative hypothesis.
(3.) State the null and alternative hypothesis from a given claim.
(4.) State the type of hypothesis test.
(5.) State the population parameter being tested.
(6.) Explain the meaning of Type I and Type II errors.
(7.) Discuss social injustice related to Type I and Type II errors. (Relate Statistics to Criminology).
(8.) Identify the Type I and Type II errors from a claim.
(9.) Identify the Type I and Type II errors in social cases. (Relate Statistics to Criminology).
(10.) Disucss the three methods used in hypothesis testing.
(11.) Test the hypothesis for a claim using the Critical Method (Classical Approach).
(12.) Test the hypothesis for a claim using the Probability Value Method (P-value Approach).
(13.) Test the hypothesis for a claim using the Confidence Interval Method.
(14.) Write the decision of the hypothesis test based on the methods used.
(15.) Write the conclusion of the hypothesis test based on the decision.
(16.) Interpret the conclusion.
(17.) Discuss the power of a hypothesis test.

Introduction

Recall:
Ask students to list the steps in the Scientific Method

Is Forming a Hypothesis one of the methods?

Is Testing the Hypothesis also one of the methods?

If you answered "yes", then you are right!

Hypothesis

Bring it to Logic: A hypothesis is defined as the premise of a conditional sattement.

Bring it to Science: A hypothesis is an unproven theory about a Scientific problem.

Bring it to Statistics: A hypothesis is a claim about a population parameter.

We are testing a claim about a population parameter using a sample statistic.
We are testing a claim about a population proportion using a sample proportion.
We are testing a claim about a population mean using a sample mean.
We are testing a claim about a population variance using a sample variance.
We are testing a claim about a population standard deviation using a sample standard deviation.

Hypothesis Test

A hypothesis test is a procedure for testing a claim about a population parameter.

Null Hypothesis

The null hypothesis is the statement that shows that the value of the population parameter is equal to some claimed value.
It is denoted by $H_0$
We test the null hypothesis assuming it to be true.
We make a decision based on the null hypothesis.
We either reject the null the hypothesis; or we do not reject the null hypothesis (or fail to reject the null hypothesis) based on the result of the methods we used for the test.
"Fail to reject the null hypothesis" is the same as "Do not reject the null hypothesis".
This does not mean that we accept the null hypothesis. It just means that we do not reject it.

Explain this concept with examples.

Then, we make a conclusion based on our decision.

Alternative Hypothesis

The alternative hypothesis is the statement that shows that the value of the population parameter is different from the claimed value.
It is denoted by $H_1$ or $H_A$ or $H_a$
The population parameter being different from the claimed value means that it could be less than the claimed value; or greater than the claimed value; or not equal to the claimed value.

If the population parameter is less than the claimed value, the hypothesis test is a left-tailed test.
In left-tailed tests, the critical region is in the extreme left region (left tail).

$population\:\:parameter \lt claimed\:\:value \implies left-tailed\:\:test$

If the population parameter is greater than the claimed value, the hypothesis test is a right-tailed test.
In right-tailed tests, the critical region is in the extreme right region (right tail).

$population\:\:parameter \gt claimed\:\:value \implies right-tailed\:\:test$

Left-tailed tests and Right-tailed tests are one-tailed tests.

If the population parameter is not equal to the claimed value, the hypothesis test is a two-tailed test.
In two-tailed tests, the critical region is in the two extreme regions (two tails: left tail and right tail).

$population\:\:parameter \ne claimed\:\:value \implies two-tailed\:\:test$

Test Statistic

The test statistic is a used in making a decision about the null hypothesis.
It is found by converting the sample statistic to a score with the assumption that the null hypothesis is true.

Methods used in Hypothesis Testing

There are three methods used in hypothesis testing.
Two of those methods will always lead to the same decision, and ultimaltely the same conclusion.
Those are the two main methods.
The other method may be used only when those two main methods are used.
The two main methods are:

(1.) Critical Value Method or Classical Approach or Traditional Method

If the test statistic is in the critical region, reject the null hypothesis.
If the test statistic is not in the critical region, do not reject the null hypothesis.

The critical values define the critical regions.
For a left-tailed hypothesis test, any value less than the critical value in the left-tail is in the critical region.
For a right-tailed hypothesis test, any value greater than the critical value in the right-tail is in the critical region.
For a two-tailed hypothesis test, any values less than the critical value in the left-tail OR greater than the critical value in the right-tail is in the critical region.

(2.) P-Value Method or Probability Value Method

If the P-value is less than or equal to the level of significance, reject the null hypothesis.
If the P-value is greater than the level of significance, do not reject the null hypothesis.

The other method is:
(3.) Confidence Interval Method

Sometimes, we use the Confidence Interval method to test the hypothesis.
We typically use this method for: one sample with two tails; two samples with left tail; two samples with right tail; or two samples with two tails.

If the confidence interval does not contain the value of the population parameter stated in the null hypothesis, reject the null hypothesis.
If the confidence interval contains the value of the population parameter stated in the null hypothesis, do not reject the null hypothesis.

For one-tailed hypothesis test: construct a confidence interval using: $CL = 1 - 2\alpha$
For two-tailed hypothesis test: construct a confidence interval using: $CL = 1 - \alpha$

NOTE:
(I.) The Classical Approach and the P-value Approach will always give the same decision and the same conclusion regardless of the hypothesis test.

(II.) For Hypothesis test about a Population Proportion; the Confidence Interval Method may or may not give the same decison and the same conclusion as the other two methods.

Explain this concept with examples.

(III.) For Hypothesis test about a: Population Mean, Population Variance, and Population Standard Deviation; all three methods will always give the same decision and the same conclusion.

What if we formed our hypothesis correctly, applied the correct methods to test it, performed our calculation correctly, and still made a wrong decision which leads to a wrong conclusion?
Is it possible?
YES. It is possible.
Why?
We are humans. We are not GOD.
Only GOD does not make mistakes.
Humans are known for making mistakes.

Errors in Hypothesis Tests

There are twi main types of errors we can make when testing hypothesis.
They are:

Type I Error (Rejecting a true null hypothesis)
This is the error made when we reject the null hypothesis when it is true.
Compare it to a False Positive scenario in Probability (Questions 50 and 51). Explain.
It is also similar to convicting an innocent man

Type II Error (Not rejecting a false null hypothesis)
This is the error made when we fail to reject the null hypothesis when it is false.
Compare it to a False Positive scenario in Probability (Questions 49 and 52). Explain.
It is also similar to acquiting a guilty man

However, which one do you think is worse?
Type I error OR Type II error?
Would you convict an innocent man? OR Would you acquit a guilty man?
Explain to students the extreme dangers of having anything to do with innocent blood.
To further explain these errors, let us review the table.

 Truth $H_0$ is true $H_0$ is false Decision Reject $H_0$ Type I error Correct decision Do not reject $H_0$ Correct decision Type II error

Compare this to: (You may use this to remember the main table)

 Truth Innocent Guilty Verdict Convict Type I error Correct decision Acquit Correct decision Type II error

$\alpha$ is the probability of making a Type I error.
$\beta$ is the probability of making a Type II error.
Type I and Type II errors are inversely related.
As $\alpha$ increases, $\beta$ decreases.
As $\alpha$ decreases, $\beta$ increases.

Level of Significance (Significance Level)

Ask students to define the level of significance (when we covered Inferential Statistics)

We are going to give another definition of the level of significance (as it concerns Hypothesis Testing).
The level of significance is the probability of making the mistake of rejecting the null hypothesis when it is true.
This implies that: The level of significance is the probability of making a Type I error.
It is denoted by $\alpha$.
NOTE: If $\alpha$ is not given, use $\alpha = 5\%$

Power of a Hypothesis Test

The power of a hypothesis test is the probability of rejecting a null hypothesis when it is false.
This implies that: The power of a hypothesis test is the probability of making the correct decision by avoiding making a Type II error.
The power of a hypothesis test depends on the significance level, the sample size, and how wrong the null hypothesis is.

Requirements for Testing Hypothesis about a Population Proportion

(1.) The samples are simple random samples.

(2.) There is a fixed number of trials.

(3.) The trials are independent.

(4.) Each trial results in either a success or a failure.

(5.) The probability of success or failure in any trial is the same as the probability of success or failure in all the trials.

(6.) There are at least ten successes and ten failures.
$np \ge 10$ AND $nq \ge 10$

(7.) The sample size is no more than five percent (at most five percent) of the population size.
$n \le 5\%N$ or $n \le 0.05N$

NOTE: If the population proportion is not given, use $50\%$
If $p$ is not given, use $p = 50\%$ or $p = 0.5$

Requirements for Testing Hypothesis about Two Population Proportions

(1.) The sample proportions are from two simple random independent samples.
Independent samples means that the sample values selected from one population proportion are not related to, or somehow natuarlly paired or matched with the sample values from the other population.

(2.) There are at least five successes and five failures for each of the two samples.
$n\hat{p} \ge 5$ AND $n\hat{q} \ge 5$ for each of the two samples.

The two main Hypothesis Test Methods uses the pooled sample proportion.
This means that the data is pooled (sort of like "pooling resources together") to obtain the proportion of successes in both samples combined.

The Confidence Interval Method uses the unpooled sample proportions.
This means the two sample proportions are treated separately.

Requirements for Testing Hypothesis about a Population Mean

(1.) The samples are simple random samples.

(2.) The population is normally distributed, OR the sample size is greater than thirty ($n \gt 30$).

Independent Samples

Two samples are independent if the sample values from one population are not related to, or somehow naturally paired or matched with the sample values from the other population.

Example: Paul was curious about the mean credit scores of men and women.
He visited the Town of Okay, Oklahoma and gathered two random samples: the credit scores of $50$ men and $50$ women.

Dependent Samples

Two samples are dependent if the sample values from one population are related to, or somehow naturally paired or matched with the sample values from the other population.
Each pair of sample values consists of two measurements from the same subject (such as before/after data), or each pair of sample values consists of matched pairs (such as husband/wife data). Example: Paul was curious about the mean credit scores of married couples.
He visited the Community of Money, Mississippi, randomly selected $50$ families, and asked for each of the credit scores of the husband and wife.

Requirements for Testing Hypothesis about Two Independent Population Means

When $\sigma_1$ and $\sigma_2$ are unknown, and are not assumed to be equal: AND
When $\mu_1$ and $\mu_2$ are assumed to be equal:
Use $t$ distribution

(1.) The values of the first population standard deviation and the second population standard deviation are unknown, and are not assumed to be equal.

(2.) The two samples are simple random samples.

(3.) The two samples are independent.

(4.) The two samples are taken from a normally distributed population, or each of the samples have sizes greater than $30$.

When $\sigma_1$ and $\sigma_2$ are unknown, but assumed to be equal:
Use $t$ distribution and a pooled sample variance

(1.) The values of the first population standard deviation and the second population standard deviation are unknown, but they are assumed to be equal.

(2.) The two samples are simple random samples.

(3.) The two samples are independent.

(4.) The two samples are taken from a normally distributed population, or each of the samples have sizes greater than $30$.

When $\sigma_1$ and $\sigma_2$ are known:
Use $z$ distribution

(1.) The values of the first population standard deviation and the second population standard deviation are known.

(2.) The two samples are simple random samples.

(3.) The two samples are independent.

(4.) The two samples are taken from a normally distributed population, or each of the samples have sizes greater than $30$.

Requirements for Testing Hypothesis about Two Dependent Population Means

Use $t$ distribution for Dependent Samples

(1.) The samples are dependent samples (matched pairs).

(2.) The two samples are simple random samples.

(3.) The two samples are taken from a normally distributed population, or each of the samples have sizes greater than $30$.

Requirements for Testing Hypothesis about a Population Variance OR a Population Standard Deviation

Use $\chi^2$ distribution

(1.) The samples are simple random samples.

(2.) The population is normally distributed.

Goodness-of-Fit Test

This is used to test the hypothesis that an observed frequency distribution fits to some claimed distribution.
It is used for analyzing categorical or quanlitative data that can be separated into different rows and columns of a table.

$H_0:$ The frequency counts agree with the claimed distribution
$H_1:$ The frequency counts do not agree with the claimed distribution.

Definitions

A hypothesis is a claim about a population parameter.

A hypothesis test is a procedure for testing a claim about a population parameter.

The null hypothesis is the statement that shows that the value of the population parameter is equal to some claimed value.

The alternative hypothesis is the statement that shows that the value of the population parameter is different from the claimed value.

A hypothesis test is a left-tailed test if the population parameter is less than the claimed value.

A hypothesis test is a right-tailed test if the population parameter is greater than the claimed value.

A hypothesis test is a two-tailed test if the population parameter is not equal than he claimed value.

The test statistic is a used in making a decision about the null hypothesis.

Type I Error is the error made when we reject the null hypothesis when it is true.

Type II Error is the error made when we fail to reject the null hypothesis when it is false.

The level of significance is the probability of making the mistake of rejecting the null hypothesis when it is true.
This implies that: The level of significance is the probability of making a Type I error.

The power of a hypothesis test is the probability of rejecting a null hypothesis when it is false.
This implies that: The power of a hypothesis test is the probability of making the correct decision by avoiding making a Type II error.

Two samples are independent if the sample values from one population are not related to, or somehow naturally paired or matched with the sample values from the other population.

Two samples are dependent if the sample values from one population are related to, or somehow naturally paired or matched with the sample values from the other population.

Symbols and Meanings

• $z_{\dfrac{\alpha}{2}}$ is the critical $z$ value
• $z_{\dfrac{\alpha}{2}}$ is the $z-score$ separating an area/probability of $\dfrac{\alpha}{2}$ in the right tail
• $-z_{\dfrac{\alpha}{2}}$ is the $z-score$ separating an area/probability of $\dfrac{\alpha}{2}$ in the left tail
• $z_{\alpha}$ is the $z-score$ separating an area/probability of $\alpha$ in the right tail
• $-z_{\alpha}$ is the $z-score$ separating an area/probability of $\alpha$ in the left tail
• $t_{\dfrac{\alpha}{2}}$ is the critical $t$ value
• $t_{\dfrac{\alpha}{2}}$ is the critical $t$ separating an area/probability of $\dfrac{\alpha}{2}$ in the right tail
• $-t_{\dfrac{\alpha}{2}}$ is the critical $t$ separating an area/probability of $\dfrac{\alpha}{2}$ in the left tail
• $z$ is the test statistic for estimating population proportion
• $z$ is the test statistic for estimating population mean (based on some conditions)
• $t$ is the test statistic for estimating population mean (based on some conditions)
• $\chi ^2$ is the test statistic for estimating population variance
• $\chi ^2$ is the test statistic for estimating population standard deviation
• $\chi ^2$ is the test statistic for Goodness-of-Fit tests
• $\alpha$ is the level of significance
• $\alpha$ is the probability of making a Type I error
• $\beta$ is the probability of making a Type II error
• $CL$ is the level of confidence
• $df$ is the degrees of freedom
• $P-value$ is the probability value
• $\hat{p}$ is the sample proportion
• $x$ is the number of individuals with the specified characteristics
• $n$ is the sample size
• $n$ is the total number of trials
• $p$ is the population proportion
• $q$ is the complement of the population proportion
• $\overline{p}$ is the pooled sample proportion
• $\overline{q}$ is the complement of the pooled sample proportion
• $s$ is the sample standard deviation
• $s^2$ is the sample variance
• $\sigma$ is the population standard deviation
• $\sigma^2$ is the population variance
• $x_1$ is the number of successes in the first sample
• $x_2$ is the number of successes in the second sample
• $\hat{p_1}$ is the first sample proportion
• $\hat{p_2}$ is the second sample proportion
• $\hat{q_1}$ is the complement of the first sample proportion
• $\hat{q_2}$ is the complement of the second sample proportion
• $\hat{p_c}$ is the critical value of the sample proportion
• $p_a$ is the alternative proportion
• $q_a$ is the complement of the alternative proportion
• $p_1$ is the first population proportion
• $p_2$ is the second population proportion
• $SE$ is the standard error
• $E$ is the margin of error
• $\overline{x_1}$ is the first sample mean
• $\overline{x_2}$ is the second sample mean
• $\overline{x_c}$ is the critical value of the sample mean
• $\overline{x_a}$ is the alternative mean
• $\mu_1$ is the first population mean
• $\mu_2$ is the second population mean
• $s_1$ is the first sample standard deviation
• $s_2$ is the second sample standard deviation
• $\sigma_1$ is the first population standard deviation
• $\sigma_2$ is the second population standard deviation
• $s_1^2$ is the first sample variance
• $s_2^2$ is the second sample variance
• $s_p^2$ is the pooled sample variance
• $\sigma_1^2$ is the first population variance
• $\sigma_2^2$ is the second population variancen
• $d$ is the individual difference between the two values in a single matched pair
• $\mu_d$ is the mean of the differences for the population of all matched pairs of data
• $\overline{d}$ is the mean value of the differences for the paired sample data
• $s_d$ is the standard deviation of the differences for the paired sample data
• $n_d$ is number of pairs of sample data
• $Ob$ is the observed frequence of an outcome (found from the sample data)
• $Ex$ is the expected frequency of an outcome (found by assuming the distribution as claimed)
• $k$ is the number of different categories
• $r$ is the sample Pearson's linear correlation coefficient
• $\rho$ is the population Pearson's linear correlation coefficient

Formulas

Proportion is given as a decimal or a percentage.
If the significance level, $\alpha$ is not given, use $\alpha = 5\%$
The null hypothesis should always have the 'equal' symbol
The alternative hypothesis has the 'unequal' symbol (less than, greater than, not equal to)

Probability of making a Type II error

To calculate the probability of making a Type II error, $\beta$
As applicable:
First Step:
Solve for the critical proportion using the critical $z$ OR
Solve for the critical mean using the critical $z$ OR
Solve for the critical mean using the critical $t$

$(1.)\:\: z_{\dfrac{\alpha}{2}} = \dfrac{\hat{p_c} - p}{\sqrt{\dfrac{pq}{n}}} \\[10ex] (2.)\:\: \hat{p_c} = z_{\dfrac{\alpha}{2}} * \sqrt{\dfrac{pq}{n}} + p \\[7ex] (3.)\:\: z_{\dfrac{\alpha}{2}} = \dfrac{\overline{x}_c - \mu}{\dfrac{\sigma}{\sqrt{n}}} \\[10ex] (4.)\:\: \overline{x_c} = z_{\dfrac{\alpha}{2}} * \dfrac{\sigma}{\sqrt{n}} + \mu \\[7ex] (5.)\:\: t_{\dfrac{\alpha}{2}} = \dfrac{\overline{x}_c - \mu}{\dfrac{s}{\sqrt{n}}} \\[10ex] (6.)\:\: \overline{x_c} = t_{\dfrac{\alpha}{2}} * \dfrac{s}{\sqrt{n}} + \mu \\[7ex]$ Second Step:
Solve for the $z$ score using the alternative proportion OR
Solve for the $z$ score using the alternative mean OR
Solve for the $t$ score using the alternative mean

$(7.)\:\: z = \dfrac{\hat{p_c} - p_a}{\sqrt{\dfrac{p_a * q_a}{n}}} \\[10ex] (8.)\:\: z = \dfrac{\overline{x}_a - \mu}{\dfrac{\sigma}{\sqrt{n}}} \\[10ex] (9.)\:\: t = \dfrac{\overline{x}_a - \mu}{\dfrac{s}{\sqrt{n}}} \\[10ex]$ Third Step:
Calculate the probability of $z$ OR
Calculate the probability of $t$

Power of a Hypothesis Test

$(1.)\:\: P_{hyp} = 1 - \beta \\[3ex]$

Hypothesis Test about a Population Proportion

If $p$ is not given, use $p = 50\%$

$(1.)\:\: \hat{p} = \dfrac{x}{n} \\[7ex] (2.)\:\: p + q = 1 \\[5ex] (3.)\:\: z = \dfrac{\hat{p} - p}{\sqrt{\dfrac{pq}{n}}} \\[10ex]$

Hypothesis Test about Two Population Proportions

The pooled sample proportion is used.
This means that the data is pooled to obtain the proportion of successess in both samples combined rather than separately (unpooled).

$(1.)\:\: \hat{p_1} = \dfrac{x_1}{n_1} \\[7ex] (2.)\:\: \hat{p_2} = \dfrac{x_2}{n_2} \\[7ex] (3.)\:\: \overline{p} = \dfrac{x_1 + x_2}{n_1 + n_2} \\[7ex] (4.)\:\: \overline{q} = 1 - \overline{p} \\[5ex] (5.)\:\: z = \dfrac{(\hat{p_1} - \hat{p_2}) - (p_1 - p_2)}{\sqrt{\dfrac{\overline{p} * \overline{q}}{n_1} + \dfrac{\overline{p} * \overline{q}}{n_2}}} \\[10ex]$

Because this is a pooled sample, we assume the null hypothesis as:

$H_0:\:\: p_1 = p_2 \:\:OR \\[3ex] H_0:\:\: p_1 - p_2 = 0 \\[3ex] \implies \\[3ex] (6.)\:\: z = \dfrac{\hat{p_1} - \hat{p_2}}{\sqrt{\dfrac{\overline{p} * \overline{q}}{n_1} + \dfrac{\overline{p} * \overline{q}}{n_2}}} \\[10ex] (7.)\:\: SE = \sqrt{\dfrac{\overline{p} * \overline{q}}{n_1} + \dfrac{\overline{p} * \overline{q}}{n_2}} \\[7ex]$

Confidence Interval Method about Two Population Proportions

The unpooled sample proportion is used.
This means that the two sample proportions are treated separately.

$(1.)\:\: E = z_{\dfrac{\alpha}{2}} * \sqrt{\dfrac{\hat{p_1} * \hat{q_1}}{n_1} + \dfrac{\hat{p_2} * \hat{q_2}}{n_2}} \\[7ex]$ The confidence interval method of the difference of the population proportions is:

$(\hat{p_1} - \hat{p_2}) - E \lt (p_1 - p_2) \lt (\hat{p_1} - \hat{p_2}) + E$

Hypothesis Test about a Population Mean

When $\sigma$ is known:

$(1.)\:\: z = \dfrac{\overline{x} - \mu}{\dfrac{\sigma}{\sqrt{n}}} \\[10ex]$ When $\sigma$ is NOT known:

$(2.)\:\: z = \dfrac{\overline{x} - \mu}{\dfrac{s}{\sqrt{n}}} \\[10ex] (3.)\:\: df = n - 1 \\[3ex]$

Hypothesis Test about Two Independent Population Means

Two samples are independent if the sample values from one population are not related to, or natuarlly paired or paired with the sample values from the other population.

When $t$ distribution is used
When $\sigma_1$ and $\sigma_2$ are unknown; and are NOT assumed to be equal

$(1.)\:\: t = \dfrac{(\overline{x_1} - \overline{x_2}) - (\mu_1 - \mu_2)}{\sqrt{\dfrac{s_1^2}{n_1} + \dfrac{s_2^2}{n_2}}} \\[10ex]$ When $\mu_1 - \mu_2$ is assumed to be $0$:

$(2.)\:\: t = \dfrac{\overline{x_1} - \overline{x_2}}{\sqrt{\dfrac{s_1^2}{n_1} + \dfrac{s_2^2}{n_2}}} \\[10ex]$ Simple Estimate of the Degrees of Freedom:

$(3.)\:\: df = the\:\:smaller\:\:value\:\:of\:\:(n_1 - 1) \:\:and\:\: (n_2 - 1) \\[3ex]$ Difficult Estimate (More Accurate Estimate) of the Degrees of Freedom:

$(4.)\:\: df = \dfrac{(A + B)^2}{\dfrac{A^2}{n_1 - 1} + \dfrac{B^2}{n_2 - 1}} \\[10ex] (5.)\:\: A = \dfrac{s_1^2}{n_1} \\[7ex] (6.)\:\: B = \dfrac{s_2^2}{n_2} \\[7ex]$ When $t$ distribution is used
When $\sigma_1$ and $\sigma_2$ are unknown; and are assumed to be equal
The pooled dample variance is used.

$(1.)\:\: t = \dfrac{(\overline{x_1} - \overline{x_2}) - (\mu_1 - \mu_2)}{\sqrt{\dfrac{s_p^2}{n_1} + \dfrac{s_p^2}{n_2}}} \\[10ex] (2.)\:\: s_p^2 = \dfrac{s_1^2(n_1 - 1) + s_2^2(n_2 - 1)}{(n_1 - 1)(n_2 - 1)} \\[7ex] (3.)\:\: s_p = \sqrt{\dfrac{s_1^2(n_1 - 1) + s_2^2(n_2 - 1)}{(n_1 - 1)(n_2 - 1)}} \\[7ex] (4.)\:\: df = n_1 + n_2 - 2 \\[3ex]$ When $z$ distribution is used
When $\sigma_1$ and $\sigma_2$ are known

$(1.)\:\: z = \dfrac{(\overline{x_1} - \overline{x_2}) - (\mu_1 - \mu_2)}{\sqrt{\dfrac{\sigma_1^2}{n_1} + \dfrac{\sigma_2^2}{n_2}}} \\[10ex]$

Confidence Interval about Two Independent Population Means

When $\sigma_1$ and $\sigma_2$ are unknown; and are NOT assumed to be equal

The confidence interval method of the difference of the population means is:

$(1.)\:\: Confidence\:\:Interval\:\:is:\:\:(\overline{x_1} - \overline{x_2}) - E \lt (\mu_1 - \mu_2) \lt (\overline{x_1} - \overline{x_2}) + E \\[5ex] (2.)\:\: E = t_{\dfrac{\alpha}{2}} * \sqrt{\dfrac{s_1^2}{n_1} + \dfrac{s_2^2}{n_2}} \\[7ex]$ Simple Estimate of the Degrees of Freedom:

$(3.)\:\: df = the\:\:smaller\:\:value\:\:of\:\:(n_1 - 1) \:\:and\:\: (n_2 - 1) \\[3ex]$ Difficult Estimate (More Accurate Estimate) of the Degrees of Freedom:

$(4.)\:\: df = \dfrac{(A + B)^2}{\dfrac{A^2}{n_1 - 1} + \dfrac{B^2}{n_2 - 1}} \\[10ex] (5.)\:\: A = \dfrac{s_1^2}{n_1} \\[7ex] (6.)\:\: B = \dfrac{s_2^2}{n_2} \\[7ex]$ When $\sigma_1$ and $\sigma_2$ are unknown; and are NOT assumed to be equal

The confidence interval method of the difference of the population means is:

$(1.)\:\: Confidence\:\:Interval\:\:is:\:\:(\overline{x_1} - \overline{x_2}) - E \lt (\mu_1 - \mu_2) \lt (\overline{x_1} - \overline{x_2}) + E \\[5ex] (2.)\:\: E = t_{\dfrac{\alpha}{2}} * \sqrt{\dfrac{s_p^2}{n_1} + \dfrac{s_p^2}{n_2}} \\[7ex] (3.)\:\: s_p^2 = \dfrac{s_1^2(n_1 - 1) + s_2^2(n_2 - 1)}{(n_1 - 1)(n_2 - 1)} \\[7ex] (4.)\:\: s_p = \sqrt{\dfrac{s_1^2(n_1 - 1) + s_2^2(n_2 - 1)}{(n_1 - 1)(n_2 - 1)}} \\[7ex] (5.)\:\: df = n_1 + n_2 - 2 \\[3ex]$ When $z$ distribution is used
When $\sigma_1$ and $\sigma_2$ are known

The confidence interval method of the difference of the population means is:

$(1.)\:\: Confidence\:\:Interval\:\:is:\:\:(\overline{x_1} - \overline{x_2}) - E \lt (\mu_1 - \mu_2) \lt (\overline{x_1} - \overline{x_2}) + E \\[5ex] (2.)\:\: E = z_{\dfrac{\alpha}{2}} * \sqrt{\dfrac{\sigma_1^2}{n_1} + \dfrac{\sigma_2^2}{n_2}} \\[7ex]$

Hypothesis Test about Two Dependent Population Means

Two samples are dependent if the sample values are somehow matched, where the matching is based on some meaningful relationship.
Each pair of sample values consists of two measurements from the same subject (such as before/after data), or each pair of sample values consists of matched pairs (such as husband/wife data).

$(1.)\:\: t = \dfrac{\overline{d} - \mu_d}{\dfrac{s_d}{\sqrt{n}}} \\[7ex] (2.)\:\: df = n - 1$

Confidence Interval about Two Dependent Population Means

The confidence interval method of the difference of the population means is:

$(1.)\:\: Confidence\:\:Interval\:\:is:\:\: \overline{d} - E \lt \mu_d \lt \overline{d} + E \\[5ex] (2.)\:\: E = t_{\dfrac{\alpha}{2}} * \dfrac{s_d}{\sqrt{n}} \\[5ex]$

Goodness-of-Fit Test

$(1.)\:\: \chi^2 = \Sigma \dfrac{(Ob - Ex)^2}{Ex} \\[7ex] (2.)\:\: df = k - 1 \\[3ex]$

Correlation

Two Samples

First Formula for the Pearson Correlation Coefficient

$(1.)\:\: r = \dfrac{\Sigma \left(\dfrac{x - \overline{x}}{s_x}\right)\left(\dfrac{y - \overline{y}}{s_y}\right)}{n - 1} \\[10ex] (2.)\:\: r = \dfrac{\Sigma(z_x)(z_y)}{n - 1} \\[5ex]$ Second Formula for the Pearson Correlation Coefficient

$(1.)\:\: r = \dfrac{n(\Sigma xy) - (\Sigma x)(\Sigma y)}{\sqrt{n(\Sigma x^2) - (\Sigma x)^2} * \sqrt{n(\Sigma y^2) - (\Sigma y)^2}} \\[7ex]$ Critical Value of the Correlation Coefficient (Use for Critical Value method)

$(1.)\:\: Critical\:\:value\:\:of\:\:r = \sqrt{\dfrac{t^2}{t^2 + df}} \\[7ex] (where\:\:t-values\:\:are\:\:from\:\:the\:\:Critical\:\:t\:\:Table) \\[3ex] (2.)\:\: df = n - 2 \\[3ex]$ Test Statistic of the Correlation Coefficient (Use for P-value method)

$(1.)\:\: t = \dfrac{r}{\sqrt{\dfrac{1 - r^2}{n - 2}}} \\[7ex]$

Decision, Conclusion, and Interpretation

NOTE:
(1.) Use $z$ or $t$ as necessary

(2.) For the interpretations, replace proportion with mean as applicable.

(3.) Use only these terms for the Decision
(a.) Reject the null hypothesis
(b.) Do not reject the null hypothesis.
(c.) Fail to reject the null hypothesis.
(b.) and (c.) means the same thing.

Keep in mind that we are dealing with the null hypothesis
You either reject the null hypothesis or you do not reject the null hypothesis.
We do not accept the null hypothesis.
We just do not reject it.
The fact that we do not reject it does not mean that we accept it.
Hence, the need for the conclusion and the interpretation.

(4.) Use only these terms for the Conclusion
(a.) There is sufficient evidence to support/warrant the rejection of the null hypothesis.
(b.) There is sufficient evidence to reject the null hypothesis.
(c.) There is sufficient evidence to support/warrant the claim of the alternative hypothesis.
(d.) There is insufficient evidence to support/warrant the rejection of the null hypothesis.
(e.) There is insufficient evidence to reject the null hypothesis.
(f.) There is insufficient evidence to support/warrant the claim of the alternative hypothesis.
(a.), (b.), and (c.) means the same thing.
(d.), (e.), and (f.) means the same thing.

IMPORTANT:
(1.) For:
Hypothesis Test about a Population Proportion OR
Hypothesis Test about a Population Mean
Use $z$ table

(2.) For:
Hypothesis Test about a Population Mean (whose Population Standard Deviation is not known)
Use $t$ table

(3.) For:
Hypothesis Test about a Population Variance OR
Hypothesis Test about a Population Standard Deviation
use $\chi^2$ table

Critical Value Method or Classical Approach

Left-tailed Tests

Find $-z_{\dfrac{\alpha}{2}}$

Condition $1$
If $z \lt -z_{\dfrac{\alpha}{2}}$, it falls in the critical region

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of the alternative hypothesis

Interpretation for One-Sample: The population proportion for the variable is significantly less than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is significantly less than the population proportion for the second variable

Condition $2$
If $z \gt -z_{\dfrac{\alpha}{2}}$, it does not fall in the critical region

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of the alternative hypothesis

Interpretation for One-Sample: The population proportion for the variable is NOT significantly less than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is NOT significantly less than the population proportion for the second variable

Right-tailed Tests

Find $z_{\dfrac{\alpha}{2}}$

Condition $1$
If $z \gt z_{\dfrac{\alpha}{2}}$, it falls in the critical region

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of the alternative hypothesis

Interpretation for One-Sample: The population proportion for the variable is significantly more than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is significantly more than the population proportion for the second variable

Condition $2$
If $z \lt z_{\dfrac{\alpha}{2}}$, it does not fall in the critical region

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of the alternative hypothesis

Interpretation for One-Sample: The population proportion for the variable is NOT significantly more than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is NOT significantly more than the population proportion for the second variable

Two-tailed Tests

Find $-z_{\dfrac{\alpha}{2}}$ and $z_{\dfrac{\alpha}{2}}$

Condition $1$
If $z \lt -z_{\dfrac{\alpha}{2}}$ OR $z \gt z_{\dfrac{\alpha}{2}}$, it falls in the critical region

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the rejection of the null hypothesis

Interpretation for One-Sample: The population proportion for the variable is significantly different than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is significantly different from the population proportion for the second variable

Condition $2$
If $-z_{\dfrac{\alpha}{2}} \lt z \lt z_{\dfrac{\alpha}{2}}$, it does not fall in the critical region

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to warrant the rejection of the null hypothesis

Interpretation for One-Sample: The population proportion for the variable is NOT significantly different than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is NOT significantly different from the population proportion for the second variable

P-Value (Probability-Value) Approach

Left-tailed Tests

Find $P(z \lt -test\:\:statistic)$

Condition $1$
If $P-value \le \alpha$

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of the alternative hypothesis

Interpretation for One-Sample: The population proportion for the variable is significantly less than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is significantly less than the population proportion for the second variable

Condition $2$
If $P-value \gt \alpha$

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of the alternative hypothesis

Interpretation for One-Sample: The population proportion for the variable is NOT significantly less than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is NOT significantly less than the population proportion for the second variable

Right-tailed Tests

Find $P(z \gt test\:\:statistic)$

Condition $1$
If $P-value \le \alpha$

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of the alternative hypothesis

Interpretation for One-Sample: The population proportion for the variable is significantly more than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is significantly more than the population proportion for the second variable

Condition $2$
If $P-value \gt \alpha$

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of the alternative hypothesis

Interpretation for One-Sample: The population proportion for the variable is NOT significantly more than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is NOT significantly more than the population proportion for the second variable

Two-tailed Tests

Find $P(z \lt -test\:\:statistic) + P(z \gt test\:\:statistic)$

Condition $1$
If $P-value \le \alpha$

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the rejection of the null hypothesis

Interpretation for One-Sample: The population proportion for the variable is significantly different than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is significantly different from the population proportion for the second variable

Condition $2$
If $P-value \gt \alpha$

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to warrant the rejection of the null hypothesis

Interpretation for One-Sample: The population proportion for the variable is NOT significantly different than the stated value.

Interpretation for Two-Samples: The population proportion for the first variable is NOT significantly different from the population proportion for the second variable

Confidence Interval Method

One Sample: Two-Tailed Tests

Construct a confidence interval using: $CL = 1 - 2\alpha$
Determine the confidence interval using an appropriate confidence level

Condition $1$
If the confidence interval does NOT the value of the population parameter stated in the null hypothesis:

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to warrant the rejection of the null hypothesis

Interpretation: The confidence interval does NOT contain the value of the population parameter stated in the null hypothesis

Condition $2$
If the confidence interval contains the value of the population parameter stated in the null hypothesis:

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of the alternative hypothesis

Interpretation: The confidence interval contains the value of the population parameter stated in the null hypothesis

Two Samples: Left-tailed Tests

Construct a confidence interval using: $CL = 1 - \alpha$
Determine the confidence interval using an appropriate confidence level

Condition $1$
If $Lower\:\:Confidence\:\:Limit \lt 0$ AND $Upper\:\:Confidence\:\:Limit \lt 0$: the confidence interval does not contain $0$

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of the alternative hypothesis

Interpretation: The population proportion for the first variable is significantly less than the population proportion for the second variable

Condition $2$
If $Lower\:\:Confidence\:\:Limit \lt 0$ AND $Upper\:\:Confidence\:\:Limit \gt 0$: the confidence interval contains $0$

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of the alternative hypothesis

Interpretation: The population proportion for the first variable is NOT significantly less than the population proportion for the second variable

Two Samples: Right-tailed Tests

Construct a confidence interval using: $CL = 1 - \alpha$
Determine the confidence interval using an appropriate confidence level

Condition $1$
If $Lower\:\:Confidence\:\:Limit \gt 0$ AND $Upper\:\:Confidence\:\:Limit \gt 0$: the confidence interval does not contain $0$

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of the alternative hypothesis

Interpretation: The population proportion for the first variable is significantly more than the population proportion for the second variable

Condition $2$
If $Lower\:\:Confidence\:\:Limit \lt 0$ AND $Upper\:\:Confidence\:\:Limit \gt 0$: the confidence interval contains $0$

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of the alternative hypothesis

Interpretation: The population proportion for the first variable is NOT significantly more than the population proportion for the second variable

Two Samples: Two-tailed Tests

Construct a confidence interval using: $CL = 1 - \alpha$
Determine the confidence interval using an appropriate confidence level

Condition $1$
If $Lower\:\:Confidence\:\:Limit \lt 0$ AND $Upper\:\:Confidence\:\:Limit \lt 0$ OR $Lower\:\:Confidence\:\:Limit \gt 0$ AND $Upper\:\:Confidence\:\:Limit \gt 0$ the confidence interval does not contain $0$

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to warrant the rejection of the null hypothesis

Interpretation: The population proportion for the first variable is significantly different from the population proportion for the second variable

Condition $2$
If $Lower\:\:Confidence\:\:Limit \lt 0$ AND $Upper\:\:Confidence\:\:Limit \gt 0$: the confidence interval contains $0$

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to warrant the rejection of the null hypothesis

Interpretation: The population proportion for the first variable is NOT significantly different from the population proportion for the second variable

Correlation: Critical Value Method or Classical Approach

Two Samples: Left-tailed Tests

Calculate the Pearson correlation coefficient
Calculate the critical value of the Pearson correlation coefficient

Condition $1$
If $r \lt 0$ AND $r \lt Critical\:\:value\:\:of\:\:r$:

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of a negative linear correlation

Condition $2$
If $r \lt 0$ AND $r \gt Critical\:\:value\:\:of\:\:r$:

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of a negative linear correlation

Two Samples: Right-tailed Tests

Calculate the Pearson correlation coefficient
Calculate the critical value of the Pearson correlation coefficient

Condition $1$
If $r \gt 0$ AND $r \gt Critical\:\:value\:\:of\:\:r$:

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of a positive linear correlation

Condition $2$
If $r \gt 0$ AND $r \lt Critical\:\:value\:\:of\:\:r$:

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of a positive linear correlation

Two Samples: Two-tailed Tests

Calculate the Pearson correlation coefficient
Calculate the critical value of the Pearson correlation coefficient

Condition $1$
If $|r| \gt Critical\:\:value\:\:of\:\:r$:

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of a linear correlation

Condition $2$
If $|r| \le Critical\:\:value\:\:of\:\:r$:

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of a linear correlation

Correlation: P-Value Method

Two Samples: Left-tailed Tests

Calculate the Pearson correlation coefficient
Calculate the test statistic
Determine the probability value (p-value) of the test statistic

Condition $1$
If $r \lt 0$ AND $P-value \le \alpha$:

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of a negative linear correlation

Condition $2$
If $r \lt 0$ AND $P-value \gt \alpha$:

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of a negative linear correlation

Two Samples: Right-tailed Tests

Calculate the Pearson correlation coefficient
Calculate the test statistic
Determine the probability value (p-value) of the test statistic

Condition $1$
If $r \gt 0$ AND $P-value \le \alpha$:

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of a positive linear correlation

Condition $2$
If $r \gt 0$ AND $P-value \gt \alpha$:

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of a positive linear correlation

Two Samples: Two-tailed Tests

Calculate the Pearson correlation coefficient
Calculate the test statistic
Determine the probability value (p-value) of the test statistic

Condition $1$
If $P-value \le \alpha$:

Decision: Reject the null hypothesis

Conclusion: There is sufficient evidence to support the claim of a linear correlation

Condition $2$
If $P-value \gt \alpha$:

Decision: Do not reject the null hypothesis

Conclusion: There is insufficient evidence to support the claim of a linear correlation

References

Chukwuemeka, S.D (2017, April 30). Samuel Chukwuemeka Tutorials - Math, Science, and Technology. Retrieved from https://www.samuelchukwuemeka.com

Gould, R., Wong, R., & Ryan, C. N. (2020). Introductory statistics : exploring the world through data. Pearson.

Kozak, K. (2014). Statistics Using Technology. ‌

Sullivan, M., Barnett, R. A., Ziegler, M. R., & Byleen, K. E. (2013). Statistics : informed decisions using data with an introduction to mathematics of finance. Pearson Learning Solutions.

Triola, M. F., & Iossi, L. (2019). Elementary statistics using the TI-83/84 plus calculator. Pearson.

Sullivan, M., & Sullivan, M. (2017). Algebra & Trigonometry ($7^{th}$ ed.). Boston: Pearson.

Weiss, N. A. (2019). Elementary statistics. Pearson. ‌

OpenStax. (n.d.). Openstax.Org. Retrieved May 16, 2018, from https://openstax.org/details/books/introductory-statistics

P(Z ≤ z-score). (n.d.). Retrieved February 24, 2019, from http://pages.stat.wisc.edu/~ifischer/Statistical_Tables/Z-distribution.pdf ‌

T Table. (2007). http://www.stat.tamu.edu/~lzhou/stat302/T-Table.pdf ‌

1.3.6.7.4. Critical Values of the Chi-Square Distribution. (n.d.). Itl.Nist.Gov. Retrieved June 17, 2020, from https://itl.nist.gov/div898/handbook/eda/section3/eda3674.htm

Radford University – Table of Critical Values for Pearson's r. (n.d.). Retrieved April 30, 2018, from https://www.radford.edu/~jaspelme/statsbook/Chapter%20files/Table_of_Critical_Values_for_r.pdf