Share

Clinical Trials

7. Statistics

Table 3

Summary of the key points from the results described in Table 2

 

Key points about Significance Test and CI Examples
In a small study, a large P-value does not mean that the null hypothesis is true – ‘absence of evidence is not evidence of absence.’ Trials 1 and 3
A large study has a better chance of detecting a given treatment effect than a small study, and is therefore more powerful. Trials 2 and 4
A small study usually produces a CI for the treatment effect that is too wide to allow any useful conclusion. Trials 1 and 3
A large study usually produces a narrow CI, and therefore a precise estimate of treatment effect. Trials 2 and 4
The smaller the P-value, the lower the chance of falsely rejecting the null hypothesis, and the stronger the evidence for rejecting the null hypothesis. Trials 2 and 4
Even if the P-value shows a statistically significant result, it does not mean that the treatment effect is clinically significant. The clinical importance of the estimated effects should always be assessed. Trial 4

CI: confidence interval.