All the following statements concerning statistical inference are true EXCEPT: (AIPG 2009)
|A||A test of statistical significance does not prove causality|
|B||A statistically significant test assesses the probability of a “chance” occurrence|
|C||A statistically significant test supports the null hypothesis|
|D||A confidence interval does not address whether an association is due to bias.|
Ref: High Yield Biostatistics, Pg: 34- 36
a. A statistically significant test (p < 0.05) rejects the null hypothesis, which states that no differences of effect will be found. A p < 0.05 indicates that a difference was found and there is <5 percent probability that it occurred by chance.
b. No test of significance can determine causality or rule out errors or bias as the cause of an association. Neither can it imply clinical significance.
c. A nonsignificant test may still have valuable clinical implications, such as the need for a larger, more involved study of a specific association.