Вы находитесь на странице: 1из 3

Lab 2

2.1 Histogram Checking Normality



Using Levenes Test for Equality of Variance to Perform an Independent Sample t-test
Group Statistics

case-control N Mean Std. Deviation Std. Error
Mean
measurement 2
control 134 82.9440 12.32271 1.06452
case 67 92.9313 9.85769 1.20431


Distribution: Normal
Mean: Standard Deviation
Control: 82.94, 12.32
Case: 92.93, 9.86

Independent sample t-test
Difference b/w two means
1
st
step in hypothesis testing: statement of study and justification of method
Statement: whether M2 is different between case and controls
Method Justification: that cases and controls are independent groups, and M2 is continuous or is measured on a
continuous scale
H0: Means of M2 for cases and controls in the population are the same
H1: means of M2 for cases and controls in the population are different
Assumptions:
- Measurements in M2 are continuous and normally distributed
- The cases and controls groups are independent
- The participants within the case and control are independent
Note: Levens test tells us whether the population variances are same or different
Decision: equal population variances as levens test p values>0.05
Step 2: calculate test statistics, t=5.77
Step 3: Sig (0.223) is >0.05 so it is not significant so variance is equal. P value<0.001. 95% CI is (6.57, 13.40) which
excludes 0 (if it included 0 then it would mean there is no difference and therefore null is true)
Step 4: Decision: Reject H0 as p-value <0.001<0.05
Step 5: Summary conclusions:
- Cases have higher mean M2 (92.93) than that of controls (82.94)
- The difference is significant (p-value<0.001, 95% CI:6.68-13.4)
- Higher value of M2 may cause the disease
Checking Normality

Using a Paired Samples t-test to Compare Before and After Systolic Blood Pressures of 40 Patients



Distribution: Normal
Mean:
Standard Deviation

Paired sample tests measure
Mean of the difference
1
st
step in hypothesis testing: statement of study and justification of method
Statement: whether the drug significantly increase the blood pressure
Method Justification: that cases and controls are not independent groups because the baseline measurements
were collected from the same group of people, and BP is continuous or is conducted on a continuous scale
H0: Means of BP before and after for cases and controls in the population are the same
H1: means of BP before and after the drug was applied to cases and controls in the population are different
Assumptions:
- The BP baseline and end of study difference are continuous and normally distributed
- The BP difference before and after are independent of each other
The 95% CI is (13.92, 24.28) which means that we are 95% confident that the true/population mean of the
difference of thebaseline BP and the end of study bP lie within 13.92 and 24.28
Step 2: calculate test statistics, t< 7.462
Step 3: Sig (0.223) is >0.05 so it is not significant so variance is equal. P value<0.001. 95% CI is (6.57, 13.40) which
excludes 0 (if it included 0 then it would mean there is no difference and therefore null is true)
Step 4: Decision: Reject H0 as p-value <0.001<0.05
Step 5: Summary conclusions:
- Mean of baselineBP (126.28) is lower than mean of end of study (145.38)
- The difference is significant (p-value<0.001, 95% CI: -13.92 to -24.28)
- The drug is effective in increasing the blood systolic pressure (directly related to the statement)
2.1.3 Compare Log_M1 between Cases and Controls Using Two Independent Samples t-test

Group Statistics

case-control N Mean Std. Deviation Std. Error Mean
Log_M1
control 134 3.5813 .39872 .03444
case 75 4.7225 .97870 .11301



Distribution: Normal
Mean: Standard Deviation
Control: 3.58, 0.399
Case:


1
st
step in hypothesis testing: statement of study and justification of method
Statement: whether the Log_M1 is different between cases and controls
Method Justification: that cases and controls are independent groups, and Log_M1 is continuous or is measured
on a continuous scale
H0: Means of Log_M1 for cases and controls in the population are the same
H1: means of Log_M1 for cases and controls in the population are different
Assumptions:
- Measurements in Log_M1 are continuous and normally distributed
- The cases and controls groups are independent
- The participants within the case and control are independent
Note: Levens test tells us whether the population variances are same or different (H0 for levens test is that the
variance of the population is the sa,mr
Decision: equal population variances cannot be assumed as levens test p values <0.001 <0.05 (therefore it is
significant) Therefore the values for equal variances not assumed should be used
Step 2: calculate test statistics, t=9.659
Step 3: Sig (<0.001) is <0.05 so it is significant so variance is not equal. P value<0.001. 95% CI is (0.906, 1.376)
which includes 0 (if it includes 0 then it means there is no difference and therefore null is true)
Step 4: Decision: Reject H0 as p-value <0.001<0.05
Step 5: Summary conclusions:
- Cases have higher mean M2 (92.93) than that of controls (82.94)
- The difference is significant (p-value<0.001, 95% CI:6.68-13.4)
- Higher value of M2 may cause the disease

Вам также может понравиться