If you read
how to design a psychological test, you may notice that you can only really develop a test for a
psychological trait that you already know about. This is just like chemistry or biology; you can't create a test for an unidentified chemical or organism. Of course, it's not all that easy to "identify" a
mental illness; at this point, mental illnesses are just collections of symptoms that may or may not have any unifying cause. It's rather like 19th century medicine. You could get treated for
catarrh, which is inflammation, sneezing, and runny nose. However, they wouldn't necessarily be able to tell you whether your symptoms were due to a minor virus, allergies, or the early stages of a serious disease--and so a test for catarrh wouldn't tell you anything about the possible cure for the problem; it would just enable you to get at the symptoms.
In the same way, the different mental illnesses officially defined in the DSM-IV are simply collections of symptoms. Anyone who exhibits certain behaviors is diagnosed with paranoid schizophrenia, but there's no evidence that all cases of paranoid schizophrenia are caused by the same thing. In fact, since we have treatments that will cure some, but not all, cases of the disease, it's pretty clear that the real causes are also various. Hopefully the 21st century will see us identify real diseases, rather than symptomatic descriptions.
For an illustration of this problem, consider the following:
A psychologist notices that some mailboxes are dysfunctional--they stop accepting mail. He notices that the dysfunctional mailboxes are often left alone by their owners. He test this theory by monitoring the length of time between owner visits and correlating it to the frequency of mailbox dysfunction--and he discovers that there is a strong correlation! He concludes that mailboxes resent being left along for long periods of time and respond by refusing to accept mail. He identifies "Mailbox Resentment Syndrome" and announces the simple cure--check your mailbox at least twice a week. No doubt his article in the Journal of Mailbox Psychology will help him get tenure...
This is a parable, not an exaggeration.