In: Statistics and Probability
To examine the effect of a drug on blood pressure, 10 diabetics before and after the blood pressures are measured and the following results are obtained. a) For this type of mass diabetics, what is the average difference between blood pressure before and after the drug? You can tell? b) Do you have an estimate of the difference between the mean and before blood pressure? c) 90% confidence interval for the mean difference in blood pressure of the mass before and after the drug Found it. d) What is the Spacing Length? How could the interval length be if there were 50 diabetics?
X(before) Y(after)
123 128
136 144
136 138
120 128
125 128
132 136
126 122
132 126
116 120
133 126
a) In this question we have two samples in which observations in one sample(Blood pressure before the drug xi ) can be paired with observations in the other sample(Blood pressure after the drug yi). So we can use a paired t-test.
The difference between blood pressure before and after the drug
di = yi − xi
The difference calculated using excel is shown below.
Y(after) |
X(before) |
Difference |
128 |
123 |
5 |
144 |
136 |
8 |
138 |
136 |
2 |
128 |
120 |
8 |
128 |
125 |
3 |
136 |
132 |
4 |
122 |
126 |
-4 |
126 |
132 |
-6 |
120 |
116 |
4 |
126 |
133 |
-7 |
Mean Difference() |
=(5+8+2+8+3+4-4-6+4-7)/10=1.7 |
Hypothesis
H0: True mean difference is zero
H1: True mean difference is not zero.
standard deviation of the difference(sd) = 5.478.
Therefore, Standard Error( ) = = = 1.732
So, we have test statistic t = mean difference/standard error = 1.7/1.732= 0.981 on 9 degress of freedom
Looking this up in t table gives p = .352086..
The result is insignificant as p > .05. Which means we cant reject the null hypothesis
So we can say that there is no evidence to show that there is difference in the blood pressure before and after the drug at 0.05 significance level.
As we were not able to reject the null hypothis, the difference between the blood pressure should be considered to be 0
b)Mean difference is 0 as per the above paired t test.
c)
confidence interval for the mean difference
t value for 90% confidence at 9 degrees of freedom= 1.833113(from t table)
Sd= 5.478, n= 10
Therefore confidence interval for = (-3.175633,3.175633)
d) Interval if there were 50 samples,
Z value for 90% confidence is 1.65
Sd= 5.478, n= 50
Therefore confidence interval for = (-1.28,1.28)