Hypothesis Testing: SUBJECT: Statistics and Probability Subject Teacher: Marilou A. Basilio

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 29

Hypothesis Testing

SUBJECT: Statistics and Probability


Subject Teacher: Marilou A. Basilio
Objectives of the Lesson:

• illustrate a statistical hypothesis


• differentiate a null hypothesis from alternative
hypothesis
• identify the concept of significance level
• identify rejection and acceptance regions
• differentiate Type I from Type II error
• Illustrate consequences of committing errors
What is a hypothesis?
Setting: Russia after the Cold War (US vs Russia) ………….see picture below
Fact: McDo is a US Company
Question: Will Russians patronize McDo considering that it is a company of their
long term rival
Hypothesis: No, they will NOT patronize it
Action: Prove it statistically……
Hypothesis?
 proposed explanation made on the basis of limited
evidence as a starting point for further investigation.

Related Terms
– conjecture, supposition,
– speculation, postulation,
– proposition, premise,
– assumption, presumption
– (Ilocano: patta-patta, Tagalog: haka-haka)
What is a statistical hypothesis?
• is an assumption about a population parameter.
This assumption may or may not be true. Use a
test statistic to test whether it is true or not.

Population Parameters: µ; (σ)


Sample Statistic: z, t, p; 2
( x , f)
Claim: Statistical Hypothesis (Statement)

Two Possible Scenarios

• Null hypothesis (claim) is true


• Null hypothesis (claim) is false
Two Possible Actions on the Statement
• reject it
• accept the statement
Example:
X= average life of Ecolum LED

Claim: lasts up to eight years

Hypothesis:
µ = 8 years
Example:
X= average number of hours
of effectiveness (skin is dry)

Claim: lasts up to 48 hours


Hypothesis: µ = 48 hours
Example:
X= average number of hours of
effectiveness (it will keep
mosquitoes away from the skin)

Claim: lasts for 4 hours


Hypothesis: µ = 4 hours
Check: Alaxan Fast Release
Advertised claim: relieves
pain in as fast as 20 minutes

1. What is the random


variable X?
2. What is the claim?
3. What is the hypothesis?
Other Examples

- A SH student claims that they spend more than 50 pesos for academic-
related expenses per day?
(to justify an increase in ‘baon’ )
Claim: more than 50 pesos spent for academic-related expenses/day.

- A SH student who claims that he/she only uses his/her phone to play ML
for 3 hours or less on the average per day?
(maybe to counter the statement of his/her Mom that he/she
has low grades because he/she is always playing the entire day)
Claim: 3 hours playing time for ML/day

- In your Entrep classes: You want to test the acceptability of a new YEMA
with crushed PINEAPPLE…
Claim: More than 50 % (proportion) of the population will like it.
Think of your own example… …. What questions
would you like to answer using hypothesis
testing?
5 Steps in Hypothesis Testing
1. State the Null Hypothesis, H0 (stated in terms of equality)
µ = 10
There is NO SIGNIFICANT DIFFERENCE

2. State the Alternative Hypothesis, H1 (stated in terms of the other claim)


µ ≠ 10 ; µ > 10; µ < 10
There is A SIGNIFICANT DIFFERENCE

3. State the significance Level, α and Critical Region (or Value)


0.05 (95% confidence level) , 0.01 (99%), 0.10 (90%)

4. Compute the value of the test statistic (z, t or p)


ഥ −µ
𝑿 ഥ −µ
𝑿 ഥ −µ
𝑿 ഥ −µ
𝑿 ෝ−𝒑𝟎
𝒑
Z= Z= σ Z= 𝒔 t= 𝒔 z=
σ ෝ𝒒
𝒑 ෝ
𝒏 𝒏 𝒏
𝒏
5. Draw a Conclusion (based on the null hypothesis)
Accept Ho or Reject Ho
Step 1: State the Null Hypothesis, H0
(in terms of the parameter)
There is no difference between the parameter
and a specific value.
Example: The owner of a factory that sells a
particular bottled fruit juice claims that the
average fill of their product is 250 ml. Is this
claim true?

Claim: Ho (Null Hypothesis): µ = 250 ml.


Step 2: State the Alternative
Hypothesis, H1 (opposite of Ho)
There is a difference between the parameter
and a specific value. (either not equal, higher or
lower than a value)

Three Ways: H1
1. No direction: use ≠ µ ≠ 250 ml
2. Less than, < µ < 250 ml
3. More than, > µ > 250 ml
Identifying the Direction of the Test
And the Rejection Region
Critical/Rejection
H1: µ ≠ 250 ml Region
-non directional – 2 tailed

−𝑡α ,v 𝑡α ,v
2 2

H1: µ > 250 ml Critical/Rejection


One tailed – right tailed Region

𝑡α,v

Critical/Rejection
H1: µ < 250 ml Region
One tailed – left tailed

−𝑡α,v
Step 3: State the significance Level, α
and Critical Region (or Value)
• α - the probability of rejecting the null hypothesis
given that it is true (a type I error)
• risk of concluding that a difference exists when
there is no actual difference

Example:
Conclusion: Mas mabilis umepekto yung Alaxan FR
kesa yung dating Alaxan. (Alaxan FR is faster in
relieving pain compared to the old Alaxan.
- There is a 5% risk that you are wrong with your
conclusion. (Is this risk acceptable?)
Step 4: Compute the value of the test
statistic (z, t or p)
ഥ −µ
𝑿
• Z= ; σ is known
σ
ഥ −µ
𝑿
• Z= σ ; σ is known and sample is used
𝒏
ഥ −µ
𝑿
• t= 𝒔 ; σ is unknown and sample is used
𝒏

ෝ−𝒑𝟎
𝒑
• z= ; sample is used
ෝ𝒒
𝒑 ෝ
𝒏
Type I and Type II Errors

β = Type II error – Ho is accepted when in fact it is false


α = Type I error – Ho is rejected when in fact it is true
Type I and Type II errors in
Hypothesis Testing
Examples of Type I and II errors
• β = Type II error
– Example: The product is a reject but is accepted
– (Customer’s Risk)

• α = Type I error
– Example: The product is acceptable but is
rejected
– (Supplier’s Risk)
Business Sector
Null Hypothesis: The product is not defective!
Type I Type II
• Ho is true: Product is • Ho is false: Product is
not defective defective
• Decision: Reject Ho • Decision: Accept Ho
• You rejected the • You accepted the
product when it is product when it is
acceptable defective (consumer’s
(supplier’s risk) risk)
Check your understanding of
Type I and Type II error
Disasters (DRRRM)
Null Hypothesis: Storm will NOT make its landfall in the
Isabela!
• Ho is false: Storm made • Ho is true: Storm did not make
its landfall in Isabela its landfall in Isabela
• Accept Ho: The person
believed in the claim (he • Reject Ho: The person in
believed the storm will Isabela did not believe in the
not make its landfall in claim (he believed the storm
Isabela) and did not will make its landfall in Isabela)
prepare for the worst
case (complacent) and prepared for the worst
case.
Result: (bad decision - Result: (good decision-better safe
disaster) than sorry)
Ex. In 2010 a storm predicted to make landfall between Isabela and Aurora actually made
its landfall in Metro Manila
Legal (ETHICS)
Null Hypothesis, H0: The man is not guilty!
• Ho is true: Man is not guilty • Ho is false: Man is guilty
(innocent).
• Accept Ho: You said the
• Reject Ho: You said the man is man is not guilty
guilty (innocent)
Result:
Result:
• Convicted an innocent person • Let go a guilty person
Consequences of a Wrong Conclusion
Ho: Drug does not have a side effect
Decision: Accept Ho (and sold the product)
Truth: Drug has significant side effects (cancer
causing)

Type II error
Correction of Error: RECALL THE DRUG (remove
in the market a specific batch of the drug)
Example: Losartan (blood pressure)
Consequences of a Wrong Conclusion
BATCH OF REINFORCED STEEL BARS (Rebars)

Ho: µ=400 Mpa (or higher)


H1: µ < 400 Mpa
Decision: Accepted Ho (and sold the batch of RSB)

Event: Building under construction using that RSB suddenly collapsed

Type II Error
Consequence: Irreversible loss
(safety concern)
Review
• How do you state a null hypothesis?
• If the alternative hypothesis has no specified
direction, is it a 2-tailed or a 1-tailed test?
• If the given significance level is 0.05 and it is a
2-tailed test, should we use tα, v or tα/2, v ?
• If Ho is true but rejected, what type of error is
committed?
Assignment:
Write it on ½ yellow pad paper to be submitted

Using your practical research topic, formulate


your claim, state your null and alternative
hypothesis and state the possible Type I and
Type II errors that you could commit.

You might also like