I still do not understand the significance of the type I error over the type II error and I do not fully understand the concept of statistical significance as defined in class on Weds. Can someone elaborate? Thanks Tanya
I think that you can look at it one of two ways. The first is that Type I error is when you think you have found an effect (due to independant variable manipulation) that doesn't really exist. In other words, although your groups are equal and your manipulation did nothing, you think it affected the results.
Type II error is when you think that your manipulation did not cause a difference between the two groups (that the means are roughly equal) when in fact, your manipulation did affect the means somewhat.
Now, you can also look at values for these types of error. Alpha for type I, and Beta for type II. I don't know where these numbers come from, but if your alpha value is too high, then you should accept the null hypothesis. If you don't a Type I error will occur. I don't really understand Beta.
I hope this helps, and that it's right. If it's not, somebody correct me. I don't want to screw anyone up!!
Ty Schepis
It is known that 5% is acceptable for a false hit. Why must this percentage be so low? Why would it make a huge difference if it was 10% so much so that you would reject the hypothesis?
Atherton, Bregnard, Burton
From class Wed. I picked up that type II error, beta, would be when you think that your two means are not due to your manipulations when in truth they reall are. Someone correct me if I am totally off base.
In response to the question about why 5% is the acceptable limit before the hypothesis is rejected: I think it would depend on your experiment and its importance. If it was an experiment in which the results would effect some ones life ( for example researching medicine) then you would probably use a percentage lower than 5% to make sure your results are very accurate. But if you were doing an experiment that wasn't so life threatening, a higher percentage would most likely be more acceptable.
In response to Burton's comment earlier on Type II Errors: A Type II Error occurs when the null hypothesis (Ho) is accepted eventhough it is false, proving the alternative hypothesis (Ha) to be true. The population means are not equal, but you fail to reject the null hypothesis anyway, so the population means are not equal do to your manipulations, but with a Type II Error you think that they are.