Hypothesis Testing

• One-Sample Tests for the Mean

– Z Test ( known)

– t Test ( unknown)

– Two-tailed and one-tailed tests

– p-value

– Connection with Confidence Interval

A statistical hypothesis is an assumption about a population parameter. This assumption may or may

not be true. 

• Hypothesis testing refers to the formal procedures used by statisticians to accept or reject

statistical hypotheses.

• Statistical Hypotheses

• The best way to determine whether a statistical hypothesis is true would be to examine the
entire population. Since that is often impractical, researchers typically examine a random
sample from the population. If sample data are not consistent with the statistical hypothesis,
the hypothesis is rejected.

Statistical Hypotheses: Ho & Ha

• The Null Hypothesis, Ho

o A suspect “no difference” claim about a population parameter

o Tested by sample statistics and either rejected or accepted based on critical value or p-
value

o Rejection of Ho implies that alternative (opposite) hypothesis is more probable

o Only the null hypothesis involves an equality, while the alternative hypothesis deals only
with inequalities

The Alternative Hypothesis, Ha or H1

o What is true when null hypothesis cannot be accepted

o Conventions with Hypothesis Statements

1. The null hypothesis is the hypothesis that is always tested.

2. The alternate hypothesis is set up as the opposite of the null hypothesis and represents the
conclusion if the null hypothesis is rejected.

3. The null hypothesis always refers to the population parameter and not the sample statistic.
 4. The statement of the null hypothesis always contains an equal sign.

5. The statement of the alternate hypothesis never contains an equal sign.

Two methods to decide if the data does or does not support Null Hypothesis:

Null Hypothesis, Ho μ = μo μ = μo, μ

Alternative Hypothesis, Ha μ ≠ μo μ < μo

1. Compare Critical value and Test statistic for chosen level of significance

Reject Ho if

Test Statistic > Critical Value

OR

-Test Statistic < -Critical Value

2. Compare P-value and Level of Significance, α

Reject Ho if P-value < α

 • Distinguishes Rejection region (associated with α) from Acceptance region, a range of values
that test statistic may take for Ho to be accepted.

• Critical value (Z, t, F, 2,…) defines boundary between regions

– Based on desired level of significance, α

Two –tailed tests

One-tailed tests
Decision Errors

Type I error - false negative

Null hypothesis rejected when in fact true

Occurs with probability 

 = level of significance - chosen!

(1- ) = confidence coefficient

Type II error - false positive

Null hypothesis accepted when in fact false

Occurs with probability 

 = consumer’s risk, rarely known

(1- ) = power of test

Depends on , difference between hypothesized and actual parameter value, and sample size

Seek proper balance between Type I and II errors

Steps of Hypothesis testing using critical value

1. State the null hypothesis, Ho, using statistical
terms.
2. State the alternate hypothesis, Ha, also using
statistical terms.
3. Choose level of significance, α. Most often
this is set at 0.05.
4. Choose sample size. Consider risks of making
Type I and Type II errors as well as budget
constraints.
5. Determine appropriate statistical technique
and corresponding test statistic.
6. Set up the critical values that divide rejection
region from non-rejection region.
7. Collect data and compute appropriate value
of test statistic.
8. Determine if test statistic falls in rejection
region or non-rejection region.
9. Make statistical decision
10. Express statistical decision in terms of
practical problem

Hypothesis Testing using p-value

• p-value is the probability of obtaining a test statistic equal to or more extreme than the result
obtained from the sample data, given that the null hypothesis is true.

Use probabilities corresponding to Z values

– If the p-value  , accept null hypothesis

– If the p-value < , reject null hypothesis

Steps in Hypothesis testing using p-value similar to that using critical value except

Step 7. Calculate p-value based on test statistic (use software!)

Step 8. Compare p-value to α

The length of a certain lumber from a national home building store is supposed to be 8.5 feet. A builder
wants to check whether the shipment of lumber she receives has a mean length different from 8.5 feet.
In this example, we are discussing a mean and therefore set up the hypotheses in terms of μ. Here the
value of μ0 is 8.5. With the builder wanting to check if the mean length is different from 8.5, she would
set up the hypotheses as a two-tailed test. what are the Hypothesis statements?

H0:μ=8.5 vs. Ha:μ≠8.5

μ = μo Null Hypothesis, Ho

μ ≠ μo Alternative Hypothesis, Ha

Z Test ( known) - Two-tailed

• Critical value (Zc) based on chosen level of significance, 

– Typically  = 0.05 (95% confidence), where Zc = 1.96 (area = 0.95/2 = 0.475)

–  = 0.01 (99%) and 0.001 (99.9%) are also common, where Z c = 2.57 and 3.29

• Null hypothesis rejected if sample Z > Z c or –Z < -Zc, where

X 
Z

n
Z Test : Example – 1

X 
Z

n
• The manufacturer claims that the mean of the ball
bearings is 0.503 inch.
• A sample of 25 ball bearings has a sample mean of
0.5018 inch. Provided that population standard
deviation of the ball bearings is 0.004. Is there
evidence that the manufacturer is not meeting the
claimed mean of 0.503 inch? Use a 0.05 significance
level).
Answer:
 ) H0 : µ = 0.503
H1 : µ ≠ 0.503
) Zc = (0.5018---X – 0.503---u)/(0.004----sigma/Sqrt----
n(25)) = -1.50

X 
Z

• n = sample mean of 0.5018 inch X -
mean of the ball bearings is 0.503 inch.

Z Test ( known) - Two-tailed

 https://www.statskingdom.com/110MeanNormal1.html

Z Test : Example – 1 (using Minitab)

X 
Z

n


https://www.statskingdom.com/130MeanT1.html - ONE TAIL