News0 min ago
level of significance
2 Answers
Hi guys
I was wondering if you could help. I am trying to understand what the level of significance is (e.g. 5%). What does it mean and how do you interpret it?
I do not understand any of the definitions in the books. Can anyone explain it as if I was a 7 year old?
I was wondering if you could help. I am trying to understand what the level of significance is (e.g. 5%). What does it mean and how do you interpret it?
I do not understand any of the definitions in the books. Can anyone explain it as if I was a 7 year old?
Answers
1 to 2 of 2
Best Answer
No best answer has yet been selected by bartholomew. Once a best answer has been selected, it will be shown here.
For more on marking an answer as the "Best Answer", please visit our FAQ.
You doing a final year project too eh?
The way I understand it is.............if you calculate your errors and they are higher than 5% then your results aren't significant and don't stand for squat as your experiments were too sloppy or you could say well they show promise but the method needs tidyign up blah blah blah. However, below 0.05/5% then you're ok, you can conclude that you don't suck :)
You can get a computer programme to do all this for you, I think I'm supposed to be using minitab but best advice, buy a statistician a cup of coffe and a buscuit and listen carefully.
The way I understand it is.............if you calculate your errors and they are higher than 5% then your results aren't significant and don't stand for squat as your experiments were too sloppy or you could say well they show promise but the method needs tidyign up blah blah blah. However, below 0.05/5% then you're ok, you can conclude that you don't suck :)
You can get a computer programme to do all this for you, I think I'm supposed to be using minitab but best advice, buy a statistician a cup of coffe and a buscuit and listen carefully.
If you conduct a hypothesis test, you define a null hypothesis - that is the condition you assume to be true unless the data suggests otherwise. You also define an alternate hypothesis which is whatever you are trying to prove.
When you carry out your test, you calculate a p-value, which is the probability that the null hypothesis is true.
Convention says that if the p-value (the level of significance) is above 5% i.e. 0.05, we haven't proved what we were trying to prove so we have to go along with the null hypothesis.
Putting it another way, if we want to show that the null hypothesis is wrong, we have to take a risk because we only work with samples. We accept a risk of 5% but no more than that. If the p-value is higher than 5% (0.05), we would be taking a greater risk than we want to if we reject the null hypothesis - so we don't.
When you carry out your test, you calculate a p-value, which is the probability that the null hypothesis is true.
Convention says that if the p-value (the level of significance) is above 5% i.e. 0.05, we haven't proved what we were trying to prove so we have to go along with the null hypothesis.
Putting it another way, if we want to show that the null hypothesis is wrong, we have to take a risk because we only work with samples. We accept a risk of 5% but no more than that. If the p-value is higher than 5% (0.05), we would be taking a greater risk than we want to if we reject the null hypothesis - so we don't.
1 to 2 of 2
