TESTING THE DIFFERENCE

Testing
Hypothesis Condition Hypothesis Critical value Test value
for:
• Use normal distribution
(𝑋̅₁ − 𝑋̅₂) − (𝜇1 − 𝜇2 )
H₁ : µ₁ ≠ µ₂ • α must divide by 2 for two tailed test 𝑧=
σ known H₁ : µ₁ > µ₂ • sign > means use right side while sign < use 𝜎2 𝜎2
H₁ : µ₁ < µ₂ left side √ 1 + 2
𝑛1 𝑛2
• α value is area under the curve (inside table)
• Use student-t distribution
• Use df = n-1
• α must divide by 2 for two tailed test (𝑋̅₁ − 𝑋̅₂) − (𝜇1 − 𝜇2 )
σ unknown; H₁ : µ₁ ≠ µ₂ 𝑡=
• sign > means use right side while sign < use
independent H₁ : µ₁ > µ₂ 𝑠12 𝑠22
left side √ +
sample H₁ : µ₁ < µ₂ 𝑛1 𝑛2
Mean; µ • t distribution differ with normal distribution
where the value inside table is the critical
value not α
• Use student-t distribution
• Use df = n-1
• α must divide by 2 for two tailed test 𝐷 = 𝑋1 − 𝑋2 ; 𝐷 ̅ = ∑𝐷
σ unknown; H₁ : µ𝐷 ≠ 0 𝑛 ̅ − 𝜇𝐷
• sign > means use right side while sign < use 𝐷
dependent H₁ : µ𝐷 > 0 𝑡=
left side 𝑠𝐷 ⁄√𝑛
sample H₁ : µ𝐷 < 0 𝑛 ∑ 𝐷 2 − (∑ 𝐷)2
• t distribution differ with normal distribution 𝑠𝐷 = √
where the value inside table is the critical 𝑛(𝑛 − 1)
value not α
• Use normal distribution 𝑋 +𝑋
H₁ : p₁ ≠ p₂ • α must divide by 2 for two tailed test 𝑝̅ = 1 2 ; 𝑞̅ = 1 − 𝑝̅ (𝑝̂1 − 𝑝̂2 ) − (𝑝1 − 𝑝2 )
𝑛1 +𝑛2
Proportion; 𝑧=
N/A H₁ : p₁ > p₂ • sign > means use right side while sign < use 1 1
p 𝑋1 𝑋2 √𝑝̅ 𝑞̅ ( + )
H₁ : p₁ < p₂ left side 𝑝
̂1 = 𝑛 ; 𝑝̂2 = 𝑛 𝑛 1 𝑛 2
• α value is area under the curve (inside table) 1 2

• use F distribution
• larger sample variance put as numerator while small sample variance put
H₁ : 𝜎12 ≠ 𝜎22
as denominator
Variance / H₁ : 𝜎12 > 𝜎22
• have two degrees of freedom; dfN, dfD (depends on numerator, 𝑠²1
Standard N/A H₁ : 𝜎12 < 𝜎22 𝐹=
denominator; formula for each df is = n-1) 𝑠²2
Deviation *same goes to
• When degree of freedom values cannot be found in the table, use the
σ
closest value on the smaller side
• Always be on right side (if only one tail use α while using two tail use α/2)
DECISION:

• Reject Hₒ if test value fall into the critical region

• Reject Hₒ if p-value less than value of α

SUMMARY

• Summarize your conclusion based on the condition/ situation of your question

• Don’t forget to insert the unit if they ask you to do so

CONFIDENCE INTERVAL:

• Means; σ known

𝜎 2 𝜎22 𝜎 2 𝜎22
(𝑋̅₁ − 𝑋̅₂) − 𝑧𝛼/2 √ 1 + < µ₁ - µ₂ < (𝑋̅₁ − 𝑋̅₂) + 𝑧𝛼/2 √𝑛1 +
𝑛 1 𝑛2 1 𝑛2

• Means; σ unknown; independent sample

𝑠 2 𝑠22 𝑠12 𝑠22

(𝑋̅₁ − 𝑋̅₂) − 𝑡𝛼 √ 1 + < µ₁ - µ₂ < (𝑋̅1 − 𝑋̅ 2 ) + 𝑡𝛼 √ +
2 𝑛1 𝑛2 2 𝑛1 𝑛2

• Means; σ unknown; dependent sample

̅ − 𝑡𝛼/2 𝑠𝐷 < 𝜇𝐷 < 𝐷

𝐷 ̅ + 𝑡𝛼/2 𝑠𝐷
𝑛√ 𝑛 √

• Proportion

̂𝑞
𝑝1̂1 ̂𝑞
𝑝2̂2 ̂𝑞
𝑝1̂1 ̂𝑞
𝑝2̂2
(𝑝
̂1 − 𝑝
̂)
2 − 𝑧𝛼/2 √ + < 𝑝1 − 𝑝2 < (𝑝
̂1 − 𝑝
̂)
2 + 𝑧𝛼/2 √ +
𝑛1 𝑛2 𝑛1 𝑛2