That's a question discussed in a recent article on the CSICOP site:
The scientific evidence for the Rorschach has always been feeble. By 1965, research psychologists had concluded that the test was useless for most purposes for which it was used. ... The Rorschach ... tends to mislabel most normal people as "sick." In addition, the test cannot detect most psychological disorders (with the exception of schizophrenia and related conditions marked by thinking disturbances), nor does it do an adequate job of detecting most personality traits.
Despite such shortcomings, the Rorschach is still administered hundreds of thousands of times each year in clinics, courts, and schools. Psychologists often use the test to help courts determine which parent should be granted custody of a child. It's used in schools to identify children's emotional problems, and in prisons to evaluate felons for parole. Convicted murderers facing the death penalty, suspected victims of sexual abuse, airline pilots suspended from their jobs for alcohol abuse--all may be given the Rorschach by a psychologist who will use the test to make critical decisions about their lives.
It sounds an awful lot like the Rorschach Test falls in the same category as some other personality tests, polygraphs, and a variety of nonsensical "tests" that use the language and aura of science in order to sell nonsense to an unsuspecting public. What's sad is the fact that information about the failure of Rorschach Tests was available as far back as the 1940s, but that hasn't stopped their popularity among some psychologists. Perhaps this is another example of people ignoring facts and science when they intrude upon beliefs that are held with serious emotional commitment.