What's the difference between significance level and confidence level?

Yes, that is the definition of confidence/significance level. Alpha is the significance level. 1 – alpha is the confidence level.

In statistical test theory, the http://en.wikipedia.org/wiki/Type_I_and_type_II_errors#Type_I_erroris the probability of rejecting the null hypothesis when it is true, which is known as Type I error. The confidence interval at confidence level [math](1-\alpha)\%[/math] for a population statistic contains all the values, when performing hypothesis test against the population statistic, should not be rejected at significance level [math]\alpha\%[/math]. So if you perform a corresponding hypothesis test while calculating confidence interval, it is correct to say that the significance level is [math]5\%[/math] if the confidence level is [math]95\%[/math].

The two concepts are closely related. In some situations, they may be exactly the same and even computed using the same statistical methods. If you carefully measure something several times or in several different people, the numbers won't be exactly the same each time. You can compute the t-statistic to get the 95% confidence limits on the measurement. This will be a range where you can be reasonably certain that the true value lies. If your average is 5 and the 95% limits are 4 to 6, you can claim that the true value is between 4 and 6. Significance levels typically come in experiments or surveys involving two or more groups. Say you want to know the effects of a drug on blood volume. You randomly assign people to a placebo or drug group and then measure blood volume. The numbers you get will vary in each group so it is isn't immediately obvious whether the difference between groups is the result of the random assignment.  So you compute a t statistic. A significance level of 5% (p < 0.05) means that there is only a 5% chance that that the difference between groups is due to chance. Technically, this isn't the same as saying that the drug has a 95% probability of being real but some people interpret the numbers this way. Here is where it gets confusing. You can calculate the averages for your drug and placebo group and then compute a confidence level for the difference. If the difference between the two averages is bigger than the 95% confidence limit then you have a significance level of 5%. This is just another way of describing the same thing. The t-statistic is computed in exactly the same way. The mechanics of computing statistics of this sort is pretty well established but interpreting these numbers isn't always easy. Keep in mind that if your drug does nothing but you run the experiment 20 times, you can expect to get one experiment showing 5% significance. Rumor has it that Psychologists have been running lots of experiments and only reporting those that produce 5% significance.