require(mosaic)
a) 66.5 to 67.0 inches.
galton = fetchData("Galton")
## Data Galton found in package.
trials = do(500) * mean(height, data = resample(galton))
## Loading required package: parallel
confint(trials)
## name lower upper level method estimate margin.of.error
## 1 result 66.53 66.99 0.95 stderr 66.76 0.2267
b) De blir olika eftersom coverage interval handlar om alla cases och confidence interval i detta fall handlar om medelvärdet av alla cases.
c) 66.1 to 67.0 inches.
trials1 = do(500) * median(height, data = resample(galton))
confint(trials1)
## name lower upper level method estimate margin.of.error
## 1 result 66.06 67.21 0.95 stderr 66.63 0.5744
a) 69.0 to 69.5 inches
trials = do(1000) * mean(height ~ sex, data = resample(galton))
qdata(c(0.025, 0.975), M, data = trials)
## quantile p
## 2.5% 69.00 0.025
## 97.5% 69.47 0.975
b) Nej.
trials = do(1000) * mean(height ~ sex, data = resample(galton))
qdata(c(0.025, 0.975), F, data = trials)
## quantile p
## 2.5% 63.90 0.025
## 97.5% 64.33 0.975
c) Det finns overlap mellan män och kvinnors individuella längder. The CIs on the means are much narrower than the distribution of individual heights.
bwplot(height ~ sex, data = galton)
a) Barely overlap.
feet = fetchData("KidsFeet")
## Data KidsFeet found in package.
mod = mm(width ~ sex, data = feet)
confint(mod)
## group 2.5 % 97.5 %
## 1 B 8.976 9.404
## 2 G 8.565 9.004
b) None overlap.
cps = fetchData("CPS85")
## Data CPS85 found in package.
mod1 = mm(wage ~ sex, data = CPS85)
confint(mod1)
## group 2.5 % 97.5 %
## 1 F 7.247 8.511
## 2 M 9.413 10.577
c) Much overlap.
mod3 = mm(wage ~ married, data = CPS85)
confint(mod3)
## group 2.5 % 97.5 %
## 1 Married 8.861 9.936
## 2 Single 7.571 9.053
d) Much overlap.
mod4 = mm(wage ~ sector, data = CPS85)
confint(mod4)
## group 2.5 % 97.5 %
## 1 clerical 6.490 8.355
## 2 const 7.448 11.556
## 3 manag 11.465 13.943
## 4 manuf 6.922 9.150
## 5 other 7.386 9.615
## 6 prof 11.051 12.844
## 7 sales 6.102 9.083
## 8 service 5.529 7.546
Range of the data = Graph C Standard deviation of the data = B Standard error of mean = A 95% confidence interval on the mean = D
sd(do(500) * with(median(length), data = resample(KidsFeet)))
## [1] 0.2374
sd(do(500) * with(sd(length), data = resample(KidsFeet)))
## [1] 0.1392
samps = do(500) * qdata(0.75, length, data = resample(KidsFeet))
sd(samps$quantile)
## [1] 0.3535
Svar: D