HW 2
HW 2
t0 = ry1 y0
S2 S2
1 2
n1 + n2
Example
(a)
H0 : 1 = 2
H1 : 1 6= 2
(b)
Z0 = ry1 y0
2 2
1 2
n1 + n2
Since Z0.025 = 1.96, which is greater than 1.35, do not reject the null.
set.seed(1112) # random order
M1 <- sample(c(16.03,16.04,16.05,16.05,16.02,16.01,15.96,15.98,16.02,15.99))
set.seed(123)
M2 <- sample(c(16.02,15.97,15.96,16.01,15.99,16.03,16.04,16.02,16.01,16))
# given population variance, even if we draw 10 sample from each group
mu1<-mean(M1)
mu2<-mean(M2)
(mu1-mu2) / sqrt(0.015^2/10 + 0.018^2/10)
1
## [1] 1.349627
(c)
p-value is 0.1771
2*pnorm(-abs(1.349627))
## [1] 0.1771357
(d)
## [1] -0.004522529
(16.015-16.005) + (1.96)*sqrt((0.015^2/10) + (0.018^2/10) ) # upper bound
## [1] 0.02452253
Example
(a)
The data is assumed to be normal since all the data points are very close to qqline.
B1 <- c(6.08,6.22,7.99,7.44,6.48,7.99,6.32,7.6,6.03,7.52)
B2 <- c(5.73,5.8,8.42,6.84,6.43,8.76,6.32,7.62,6.59,7.67)
diff <- B1-B2
qqnorm(diff);qqline(diff,col=2)
2
Normal QQ Plot
0.4
Sample Quantiles
0.0
0.4
0.8
Theoretical Quantiles
(b)
The 95% confidence interval on the difference in mean score is (-0.3664148, 0.2644148) contains the value of
zero. There is no difference in birth order.
var.test(B1,B2)
##
## F test to compare two variances
##
## data: B1 and B2
## F = 0.59148, num df = 9, denom df = 9, p-value = 0.4461
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
## 0.1469164 2.3813137
## sample estimates:
## ratio of variances
## 0.5914846
t.test(diff,alternative=c("two.sided"),cond.level=0.95,var.equal = TRUE)
##
## One Sample t-test
##
## data: diff
## t = -0.36577, df = 9, p-value = 0.723
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## -0.3664148 0.2644148
## sample estimates:
3
## mean of x
## -0.051
Example
(a)
For both F1 and F2, most of the data points are closely to the qqline. Hence we can assume that the
assumption of normality is fulfilled. In addition, since the slopes for both plot are almost the same, we can
say that the assumption of equal variance is fulfilled for both samples. In addition, p-value for equal variance
is 0.9419, hence we do not reject the null that their variances are different.
set.seed(100) # random order
F1<-sample(c(206,188,205,187,193,207,185,189,192,210,194,178))
set.seed(1111)
F2<-sample(c(177,197,206,201,176,185,200,197,198,188,189,203))
qqnorm(F1);qqline(F1,col=2)
Normal QQ Plot
210
Sample Quantiles
200
190
180
Theoretical Quantiles
qqnorm(F2);qqline(F2,col=2)
4
Normal QQ Plot
205
Sample Quantiles
195
185
175
Theoretical Quantiles
qqnorm(F1-F2);qqline(F1-F2,col=2)
Normal QQ Plot
30
20
Sample Quantiles
10
0
10
20
Theoretical Quantiles
var.test(F1,F2)
##
## F test to compare two variances
5
##
## data: F1 and F2
## F = 1.046, num df = 11, denom df = 11, p-value = 0.9419
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
## 0.3011181 3.6334674
## sample estimates:
## ratio of variances
## 1.045994
(b)
Since p-value is 0.3667, which is greater than 0.05, do not reject the null. Hence the mean deflection
temperature under load for formulation 1 does not exceed that of formulation 2.
t.test(F1,F2,alternative=c("greater"),cond.level=0.95,var.equal = T)
##
## Two Sample t-test
##
## data: F1 and F2
## t = 0.34483, df = 22, p-value = 0.3667
## alternative hypothesis: true difference in means is greater than 0
## 95 percent confidence interval:
## -5.637883 Inf
## sample estimates:
## mean of x mean of y
## 194.5000 193.0833
(c)
p-value for equal variance is 0.9419, hence we do not reject the null that their variances are different. The
p-value for normal test on F1 is 0.3438 and the p-value for normal test on F2 is 0.2453. Since both p-values are
greater than 0.05, we do not reject the null hypothesis that they follow an approximately normal distribution.
shapiro.test(F1)
##
## Shapiro-Wilk normality test
##
## data: F1
## W = 0.92643, p-value = 0.3438
shapiro.test(F2)
##
## Shapiro-Wilk normality test
##
## data: F2
## W = 0.91474, p-value = 0.2453
shapiro.test(F1-F2)
##
## Shapiro-Wilk normality test
6
##
## data: F1 - F2
## W = 0.98109, p-value = 0.9876
Example
(a)
H0 : 1 = 2
H1 : 1 6= 2
(b)
Since t0.975,18 = 2.1, which is greater than 0.798, do not reject the null.
set.seed(1112) # random order
M1 <- sample(c(16.03,16.04,16.05,16.05,16.02,16.01,15.96,15.98,16.02,15.99))
set.seed(123)
M2 <- sample(c(16.02,15.97,15.96,16.01,15.99,16.03,16.04,16.02,16.01,16))
mu1<-mean(M1)
mu2<-mean(M2)
sd1<-sd(M1)
sd2<-sd(M2)
spool<-sqrt((9*sd1^2+9*sd2^2) / (18))
t0 = (mu1-mu2) / (spool * sqrt(1/10 + 1/10))
t0
## [1] 0.7989355
qt(0.975,18) # df = n1+n2-2
## [1] 2.100922
(c)
p-value is 0.43474
(1-pt(t0,18))*2
## [1] 0.4347438
t.test(M1,M2,alternative=c("two.sided"),cond.level=0.95,var.equal = T)
##
## Two Sample t-test
##
## data: M1 and M2
## t = 0.79894, df = 18, p-value = 0.4347
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.01629652 0.03629652
## sample estimates:
## mean of x mean of y
7
## 16.015 16.005
(d)
## [1] -0.01629616
(16.015-16.005) + (2.1) * 0.028 * sqrt((1/10 + 1/10))
## [1] 0.03629616
N1 <- sample(c(6625,6000,5450,5200,5175,4900,4750,4500,3985,900,450,2800))
N2 <- sample(c(3900,3500,3450,3200,2980,2800,2500,2400,2200,1200,1150,1130))
var.test(N1,N2)
##
## F test to compare two variances
##
## data: N1 and N2
## F = 3.9874, num df = 11, denom df = 11, p-value = 0.03052
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
## 1.147892 13.851139
## sample estimates:
## ratio of variances
## 3.987432
t.test(F1,F2,alternative=c("two.sided"),cond.level=0.95,var.equal = F)
##
## Welch Two Sample t-test
##
## data: F1 and F2
## t = 0.34483, df = 21.989, p-value = 0.7335
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -7.103683 9.937016
## sample estimates:
## mean of x mean of y
## 194.5000 193.0833