3.
Estimate of population proportion of horses that requires surgery is given by 24 = 0.48 (b) We have to test for null hypothesis Ho: p = 0.5 against the alternative hypothesis H1 :p < 0.5 (c) We have to perform test of proportion for one population case. (d) We have, level of significance a = 0.05 Our decision rule is "we reject our null hypothesis if p-value <0".
(e) Our test statistic is given by - po po (1 - po) /n Here, p = 0.48 po = 0.5 Sample size n = 50 ulated 0.48 – 0.5 == -0.2828427 0.5(1 – 0.5) /50 p-value = P(:< -0.2828427) = 0.3886487 [Using R-code 'pnorm(-0.2828427)]
The R calculation is as follows. >n<-50 > r<-24 > p_hat<-r/n > P_O<-0.5 > Z<-(p_hat-p_0)/sqrt (p_0* (1-p_0)/n) > Z [1] -0.2828427 > p<-pnorm (z) 2 [1] 0.3886487 We cannot reject our null hypothesis. We reject our null hypothesis if p-value <a But here we observe that p-value = 0.3886487 0.05 = a
(g) Based on the given data we can conclude that there is no significant evidence that less than half of the horses that show colic symptoms require surgery. (h) We used normal approximation of binomial distribution which is valid if • Number of success 10 • Number of failure 10 Considering having surgery as success we have • Number of success = 24 > 10 • Number of failure = 26 > 10 So, assumption for the hypothesis test is satisfied.
We know, P(-1.644854<:<1.644854) = 0.90 [Using R-code 'qnorm(1-1-0.90)/2)] + P(-1.691851 < p ="BTH < 1.61854) = 0,00 =P(-1.641854 < 70.8415 0.)750 < 1.611854) = 0,90 = P(-1.644854 < 0.04865.205 < 1.644854) = 0.90 - 1 54 <- 0.48 - P 0.48 (1 – 0.48)/50 .644 1 = 0.90 →P(-1.644854 * 0.07065409 < 0.48-P < 1.644854 * 0.07065409) = 0.90 =P(0.48 – 0.1162157 < p < 0.48 +0.1162157) = 0.90 :P(0.3637843<p < 0.5962157) = 0.90 Hence, our calculated 90% confidence interval for the population proportion of horses that require surgery is given by (0.3637843, 0.5962157).