StatisticsThreshold of p-value: If p<0.05, then there is significant effect. For p- values, p = 0.04 and p = 0.06, p= 0.04 is significant while p=0.06 is not. Similarly there is not much difference between p=0.0499 and p=0.501 and has same effect. Now alpha values whether .05 or .1 is basically an arbitrary threshold and thus saying it statistically significant or not doesn’t provide us with tr...
Threshold of p-value: If p<0.05, then there is significant effect. For p- values, p = 0.04 and p = 0.06, p= 0.04 is significant while p=0.06 is not. Similarly there is not much difference between p=0.0499 and p=0.501 and has same effect. Now alpha values whether .05 or .1 is basically an arbitrary threshold and thus saying it statistically significant or not doesn’t provide us with true scenario.
Another criticism that NHST has is that the test largely depends on the sample size. For n=10, after apply t test, the result might not be significant. But if I draw 100 samples for the same problem, and apply t test, then I might find a significant difference. Thus with the increment in sample size, there is still a significant difference even though very little.
Statistically-significant results should not be over rated and non-statistically-significant results should not be underrated. In fact results should be interpreted carefully. Even if a result is significant, it is “significant” only in the context of statistics, and it does not necessarily mean that it is meaningful or important in a practical meaning.
Login | Sign Up
