Worksheet 23 1 Hypothesis testing, Slides of Statistics

Let the null hypothesis be that the patient's blood pressure is µ = 120, and the alternative hypothesis H1 : µ > 120 with.

Typology: Slides

2021/2022

Uploaded on 08/01/2022

hal_s95
hal_s95 🇵🇭

4.4

(655)

10K documents

1 / 4

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Week 13 Worksheet 23 MATH 10B
Thu 4/18/19
1 Hypothesis testing
1. Let Xbe a measurement of a patient’s blood pressure, which follows the normal distribution
with σ= 5 mmHg center around the true blood pressure. Let the null hypothesis be that
the patient’s blood pressure is µ= 120, and the alternative hypothesis H1:µ > 120 with
significance level α= 0.05? Supposed that the patient’s blood pressure is measure to be 135
mmHg.
(a) Is it a one-sided test?
(b) Calculate the p-value.
(c) Calculate critical value and find the rejection region.
(d) Draw a conclusion.
Solution:
i. It is a one sided test.
ii. Assume the null is true, z=Xµ
σ=135120
5= 3. Looking up the table, Three
SD corresponds to an area of 0.4987. Since it is a one-sided test, the p-value is
0.50.4987 0.0013.
iii. Let kbe the critical value. Then, assuming the null is true, P(X > k) = 0.05 =
P(X120
5>k120
5) = P(Z > k120
5) = 0.05. Thus, k120
5>1.65 =k > 128.25,
and the rejection region is [128.25,).
iv. Since the p-value is 0.0013 <0.05 or X= 135 >128.25, we can reject the null and
conclude that µ > 120.
2. Let Xbe a measurement of a patient’s blood pressure, which follows the normal distribution
with σ= 5 mmHg center around the true blood pressure. Let the null hypothesis be that
the patient’s blood pressure is µ= 120, and the alternative hypothesis H1:µ6= 120 with
significance level α= 0.05? Supposed that the patient’s blood pressure is measure to be 130
mmHg.
(a) Is it a one-sided test?
(b) Calculate the p-value.
(c) Draw a conclusion.
Solution:
i. It is a two sided test.
ii. Assume the null is true, z=Xµ
σ=130120
5= 2. Looking up the table, Three
SD corresponds to an area of 0.4772. Since it is a one-sided test, the p-value is
2(0.50.4772) 0.0456.
iii. Since the p-value is 0.0456 <0.05, we can reject the null and conclude that µ6= 120.
3. You have a coin that you suspect is biased toward coming up heads. You let pbe the
probability that the coin will give heads on any one flip. Can you reject the null hypothesis
H0:p=1
2in favor of the alternative hypothesis H1:p > 1
2with significance level α= 0.05?
pf3
pf4

Partial preview of the text

Download Worksheet 23 1 Hypothesis testing and more Slides Statistics in PDF only on Docsity!

Worksheet 23

Thu 4/18/

1 Hypothesis testing

  1. Let X be a measurement of a patient’s blood pressure, which follows the normal distribution with σ = 5 mmHg center around the true blood pressure. Let the null hypothesis be that the patient’s blood pressure is μ = 120, and the alternative hypothesis H 1 : μ > 120 with significance level α = 0.05? Supposed that the patient’s blood pressure is measure to be 135 mmHg. (a) Is it a one-sided test? (b) Calculate the p-value. (c) Calculate critical value and find the rejection region. (d) Draw a conclusion. Solution: i. It is a one sided test. ii. Assume the null is true, z = X−σ μ = 135 − 5 120 = 3. Looking up the table, Three SD corresponds to an area of 0.4987. Since it is a one-sided test, the p-value is
  2. 5 − 0. 4987 ≈ 0 .0013. iii. Let k be the critical value. Then, assuming the null is true, P (X > k) = 0. 05 =⇒ P (X− 5120 > k− 5120 ) = P (Z > k− 5120 ) = 0.05. Thus, k− 5120 > 1. 65 =⇒ k > 128 .25, and the rejection region is [128. 25 , ∞). iv. Since the p-value is 0. 0013 < 0 .05 or X = 135 > 128 .25, we can reject the null and conclude that μ > 120.
  3. Let X be a measurement of a patient’s blood pressure, which follows the normal distribution with σ = 5 mmHg center around the true blood pressure. Let the null hypothesis be that the patient’s blood pressure is μ = 120, and the alternative hypothesis H 1 : μ 6 = 120 with significance level α = 0.05? Supposed that the patient’s blood pressure is measure to be 130 mmHg. (a) Is it a one-sided test? (b) Calculate the p-value. (c) Draw a conclusion. Solution: i. It is a two sided test. ii. Assume the null is true, z = X−σ μ = 130 − 5 120 = 2. Looking up the table, Three SD corresponds to an area of 0.4772. Since it is a one-sided test, the p-value is 2(0. 5 − 0 .4772) ≈ 0 .0456. iii. Since the p-value is 0. 0456 < 0 .05, we can reject the null and conclude that μ 6 = 120.
  4. You have a coin that you suspect is biased toward coming up heads. You let p be the probability that the coin will give heads on any one flip. Can you reject the null hypothesis H 0 : p = 12 in favor of the alternative hypothesis H 1 : p > 12 with significance level α = 0.05?

Worksheet 23

Thu 4/18/

(a) you get 7 heads in 8 flips Solution: Assuming the null is true, the p−value in this case is

8

(^12 )^7 (1− 12 )+

8

(^12 )^8 =

  1. 03515625 < 0 .05. Thus, reject the null.

(b) you get 6 heads in 8 flips

Solution: Assuming the null is true, the p − value in this case is

8

( (^12 )^6 (1^ −^12 )^2 +

7 8

(^12 )^7 (1 − 12 ) +

8

(^12 )^8 = 0. 14453125 > 0 .05. Thus, not sufficient evidence to reject the null.

  1. You have a coin that you suspect is biased toward coming up heads. You let p be the probability that the coin will give heads on any one flip.

(a) Explain why, if you flip the coin three times, then no matter how the three flips turn out, you will NOT have enough evidence to reject the null hypothesis “p = 12 ” in favor of the alternative hypothesis “p > 12 at significance level α = 0.05. Hint: calclute the p − values if you have 3, 2 , 1 , 0 heads and compare with α.

Worksheet 23

Thu 4/18/

Solution: If H 0 were true (i.e. p =

), for H 0 to be true and Ha to be false, there is nothing more extreme than flipping a coin n times and getting n heads. So if one flips a coin n times and get n heads, the p-value for your data is:

P (X ≥ n) = P (X = n) =

n n

)n( 1 −

)n−n = 1 ·

2 n^

2 n^

We would like to know the smallest positive integer n for which

2 n^ < 0 .05. If n = 4,

1 2 n^

but if n = 5, 1 2 n^

So 5 is the minimum number of flips one needs to reject “p =

”” in favor of “p >

at a significance level of α = 0.05.