The generalizability theory is truly reliable. I think to general the ability to think things it makes you reliable in the fact that what you are thinking is not just an opinion but based on facts that you have learned through time and from other influences that you experiance day by day.
Yes, the generalizability theory is truly reliable, because it is one common way to analyze statistics. The theory is one way of breaking down the data and seeing what it means in principle. It is also to consider what a policy or other plan does not impact the group as a whole. This is an effective way to look at it.
No, it's not the end all, be all when it comes to reliability, but it can be used wisely. What you have to be weary of, for instance, is using it to examine teachers or something like that. For the most part though, it's rather reliable. General theories do typically work however sad that fact is.
I believe the generalizability theory can be useful when it is used as one of many methods to measure effectiveness. It is not fully reliable on its own but it can be used in a variety of different tests. I think generalizability can be useful in a variety of fields; education and government being two major categories.