0% found this document useful (0 votes)
4 views13 pages

Statistical Analysis of Hypothesis Testing

The document outlines various statistical analyses including confidence intervals, hypothesis testing, and p-value calculations for different sample sizes and methods. It demonstrates the use of both z-tests and t-tests depending on sample size and variance assumptions, concluding whether to reject or not reject null hypotheses based on p-values. Additionally, it includes a chi-squared test for categorical data comparison between two methods.

Uploaded by

yentrangpham0310
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as XLSX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
4 views13 pages

Statistical Analysis of Hypothesis Testing

The document outlines various statistical analyses including confidence intervals, hypothesis testing, and p-value calculations for different sample sizes and methods. It demonstrates the use of both z-tests and t-tests depending on sample size and variance assumptions, concluding whether to reject or not reject null hypotheses based on p-values. Additionally, it includes a chi-squared test for categorical data comparison between two methods.

Uploaded by

yentrangpham0310
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as XLSX, PDF, TXT or read online on Scribd

Sample size 100 > 30 => Standard normal distribution as the sampling distribution

Alpha 0.05
Mean 75.00
Standard deviation 30.00
Confidence interval 5.88 =[Link](B2,B4,B1)
Confidence lower limit 69.12 =B3 - B5
Confidence upper limit 80.88 =B3 + B5

Confidence Interval is 69.12 to 80.88


Sample size 16 <30 => Use t-distribution as the sampling distribution of the mean
Alpha 0.05
Mean 75.00
Standard deviation 30.00
Confidence interval 15.99 =CONFIDENCE.T(B2,B4,B1)
Confidence lower limit 59.01 =B3-B5
Confidence upper limit 90.99 =B3+B5

Confidence Interval is 59.01 to 90.99


Our Data Sample size 256
Mean 102.50
Population standard deviation 16
Step 1: Set up hypothesis
Null hypothesis μ = 100
Alternative hypothesis μ ≠ 100
Step 2: Because the population variance is known => z-test
Step 3: Determine alpha and tailed test
Alpha 0.05
Compare μ <> 100 => two-tailed test
Step 4: Test statistic 2.500 =(C2-C10)/(C3/ SQRT(C1))
p-value (left-tailed test) 0.994 =[Link]( C11, TRUE())
p-value (right-tailed test) 0.006 =1 - C12
p-value (two-tailed test) 0.012 =2 * MIN(C12:C13)
Because p-value of two-tailed test < Alpha, we reject the Null hypothesis
This means that there is enough evidence to claim the students in the high school have
Step 5: a different IQ score than 100.
Our Data Sample size 64
Mean 102.50
Population standard deviation 16
Step 1: Set up hypothesis
Null hypothesis μ≤100
Alternative hypothesis μ>100
Step 2: Because the population variance is known
Step 3: Determine alpha and tailed test
Alpha 0.05
Compare μ > 100
Step 4: Test statistic 1.250
p-value (left-tailed test) 0.894
p-value (right-tailed test) 0.106
Because p-value of two-tailed test > Alpha, we do not reject the Null hypothesis
This means that there is not enough evidence to claim the students in the high school have
Step 5: an IQ score greater than the average IQ of 100.
=> z-test

=> right-tailed test


=(C2-C10)/(C3/ SQRT(C1))
=[Link]( C11, TRUE())
=1 - C12
do not reject the Null hypothesis
claim the students in the high school have
Our Data Sample size 20
Mean 102.50
Sample standard deviation 16
Step 1: Set up hypothesis
Null hypothesis μ≤100
Alternative hypothesis μ>100
Because the population variance is unknown and the
Step 2: sample size is < 30 => t-test
Step 3: Determine alpha and tailed test
Alpha 0.05
Compare μ > 100 => right-tailed test
Step 4: Test statistic 0.699 =(C2-C10)/(C3/ SQRT(C1))
Degree of freedom 19 =C1-1
p-value (left-tailed test) 0.753 =[Link]( C11, C12,TRUE())
p-value (right-tailed test) 0.247 =1 - C13
Because p-value of two-tailed test > Alpha, we do not reject the Null hypothesis
This means that there is NOT enough evidence to claim the students in the high school
Step 5: have an IQ score greater than the average IQ of 100.
Our Data Sample size 1 (New method) 32
Mean sample 1 103
Standard deviation of population 1 16
Sample size 2 (Old method) 32
Mean sample 2 100.8
Standard deviation of population 2 16
Step 1: Set up hypothesis
Null hypothesis μNew − μOld ≤ 0
Alternative hypothesis μNew − μOld >0
Step 2: Assume both population variances are known
Step 3: Determine alpha and tailed test
Alpha 0.05
Compare μNew − μOld > 0
Step 4: Standard error 4.000
Test statistic 0.550
p-value (left-tailed test) 0.709
p-value (right-tailed test) 0.291
Step 5: Because p-value of right-tailed test > Alpha, we do not reject the Null hypothesis
This means that there is not enough evidence to claim the new method is better than the old one.
=> z-test

=> right-tailed test


=SQRT(C3*C3/C1 + C6*C6/C4)
=((C2 - C5 ) - C13)/C14
=[Link]( C15, TRUE())
=1 - C16
ot reject the Null hypothesis
he new method is better than the old one.
Our Data Sample size 1 (New method) 32
Mean sample 1 103
Standard deviation of sample 1 16
Sample size 2 (Old method) 32
Mean sample 2 100.8
Standard deviation of sample 2 14
Step 1: Set up hypothesis
Null hypothesis μNew − μOld ≤ 0
Alternative hypothesis μNew − μOld >0
Step 2: Assume both population variances are unknown but equal
Step 3: Determine alpha and tailed test
Alpha 0.05
Compare μNew − μOld > 0
Step 4 Degree of freedom 62
Pooled variance 226.000
Standard error 3.758
Test statistic 0.585
p-value (left-tailed test) 0.720
p-value (right-tailed test) 0.280
Step 5: Because p-value of right-tailed test > Alpha, we do not reject the Null hypothesis
This means that there is not enough evidence to claim the new method is better than the old one.
=> t-test

=> right-tailed test


=C1 + C4 -2
=((C1-1) * C3*C3 + (C4-1) *C6*C6) / C14
=SQRT(C15) * SQRT(1/C1 + 1/C4)
=((C2 - C5 ) - C13)/C16
=[Link]( C17, C14, TRUE())
=1 - C18
ot reject the Null hypothesis
e new method is better than the old one.
Our Data: Sample size 1 10
Mean sample 1 87
Standard deviation of sample 1 14
Sample size 2 12
Mean sample 2 99
Standard deviation of sample 2 18
Step 1: Set up hypothesis
Null hypothesis μA − μB = 0
Alternative hypothesis μA − μB ≠ 0
Step 2: Assume both population variances are unknown but unequal
Step 3: Determine alpha and tailed test
Alpha 0.05
Compare mean 1 - mean 2 <> 0

Solve: Degree of freedom 20


Standard error 6.826
t statistic -1.758
p-value (left-tailed test) 0.047
p-value (right-tailed test) 0.953
p-value (two-tailed test) 0.095
Step 5: Because p-value of two-tailed test > Alpha, we do not reject the Null hypothesis
This means that there is not enough evidence to claim the new method is better than the old one.
=> t-test

=> two-tailed test

=POWER((C3*C3/C1 + C6*C6/C4),2) /
( 1/(C1-1) * POWER(C3*C3/C1,2) +
1/(C4-1) * POWER(C6*C6/C4,2) )
=SQRT( C3*C3/C1 + C6*C6/C4)
=((C2 - C5 ) - C13)/ C15
=[Link]( C16, C14, TRUE())
=1 - C17
=2 * MIN(C17:C18)
t reject the Null hypothesis
he new method is better than the old one.
Observed data:
Pass Fail Total
Method A 14 38 52
Method B 20 10 30
Total 34 48 82

Step 1: Calculate Expected Value (The percentage of each category)


Pass Fail
Method A 21.56 30.44 =C5 * E3 / E5 =D5 * E3 / E5
Method B 12.44 17.56 =C5 * E4 / E5 =D5 * E4 /E5

Step 2: Calculate p-value


P-value 0.000433746 =[Link](C3:D4, C9:D10)

Step 3: If p-value < a => Reject null hypothesis; otherwise, do not reject
Conclusion: 0.000433746 < 0.05 => Reject the null hypothesis

You might also like