In: Statistics and Probability
Consider the hypothesis test with null hypothesis μ1=μ2 and alternative hypothesis μ1 > μ2.Suppose that sample sizes n1 = 10 and n2 =10, that the sample means are 4.9 and 2.8 respectively, and that the sample variances are 2 and 3 respectively. Assume that the population variances are equal and that the data are drawn from normal distributions.
a) Test the hypothesis that at α= 0.05 and provide a conclusion statement
b) Provide an adequate confidence interval with 95% confidence level. c) Do your answers in a) and b) coincide? Explain.
a)
Ho : µ1 - µ2 = 0
Ha : µ1-µ2 > 0
Level of Significance , α =
0.05
Sample #1 ----> 1
mean of sample 1, x̅1= 4.90
standard deviation of sample 1, s1 =
1.41
size of sample 1, n1= 10
Sample #2 ----> 2
mean of sample 2, x̅2= 2.80
standard deviation of sample 2, s2 =
1.73
size of sample 2, n2= 10
difference in sample means = x̅1-x̅2 =
4.9000 - 2.8 =
2.10
pooled std dev , Sp= √([(n1 - 1)s1² + (n2 -
1)s2²]/(n1+n2-2)) = 1.5811
std error , SE = Sp*√(1/n1+1/n2) =
0.7071
t-statistic = ((x̅1-x̅2)-µd)/SE = (
2.1000 - 0 ) /
0.71 = 2.9698
Degree of freedom, DF= n1+n2-2 =
18
t-critical value , t* =
1.734 (excel function: =t.inv(α,df)
Decision: | t-stat | > | critical value |, so,
Reject Ho
p-value =
0.0041 [excel function: =T.DIST.RT(t stat,df)
]
Conclusion: p-value <α , Reject null
hypothesis
There is enough evidence to that μ1 > μ2
b)
Degree of freedom, DF= n1+n2-2 =
18
t-critical value = t α/2 =
2.1009 (excel formula =t.inv(α/2,df)
pooled std dev , Sp= √([(n1 - 1)s1² + (n2 -
1)s2²]/(n1+n2-2)) = 1.5811
std error , SE = Sp*√(1/n1+1/n2) =
0.7071
margin of error, E = t*SE = 2.1009
* 0.71 = 1.49
difference of means = x̅1-x̅2 =
4.9000 - 2.800 =
2.1000
confidence interval is
Interval Lower Limit= (x̅1-x̅2) - E =
2.1000 - 1.4856 =
0.6144
Interval Upper Limit= (x̅1-x̅2) + E =
2.1000 + 1.4856 =
3.5856
since, both ends of CI is positive, so, reject Ho
yes, both answers coincide