I would not recommend an independent t-test, because your two 'groups' are not really independent. There is an one-sample t-test that compares the empirical mean of a sample with an user-defined mean. This could be a good way to check if the mean of your sample is equal to the population mean (provided that you know the population mean).

Thanks a lot. Actually the entire group gives the population mean, since all children in Norway registered with ADHD (who are the ones I'd like to compare with) is included. 

My knowledge within statistics is not very good, so I tried to investigate whether bootstrapping could be a method to reduce the risk of making a type-I-error. But I am not sure it that is the way to og...?

Ok, that is interesting. I did not know that there is somekind of database where all children with ADHD in Norway are registered.

I am not sure about the usefulness of bootstrapping in this case. This is often helpful if the samples are very small. In your case I would recommend the one-sample t-test. Due to the big size of the whole sample (N = 11.000) a significant difference betwen the two means is very likely. You should in addition estimate the effect size to evaluate the meaningfulness of the difference.

Thanks a lot, Brian. I have run the statistics and compared the one-sample t-test to the independent t-test method. The findings are very similar, but the CI is broader for the one-sample t-test (at the third decimal Level, though). I will use the method you suggested for the analyses, thanks again.

.