The so-called Lüscher-Farbtest is not unknown in psychology, though rarely used in institutions and settings needing to take care for methodics, because as fas as I know there is little statistical work done to verify it's validity. However, it gets used by some therapists not working in har-core clinical contexts, but more ion the advisory, councelling field, and in the job assessment setting. I used it myself several times, and for reasons of curiosity many of us students back then did: we experimented, like students do.
Surprisingly I found it to have a very high hit quota.
Lüscher has chnaged the test several times in an attempt to reduce it to the very basic needs, he reduced the number of colour cards and the number of steps that were needed to do. So, several versions of the Lüscher-Test are in circulation. I prefer the one where in four trials four colours must be sorted in sequence of preference. There are shorter and longer test versions.
The general problem I have with diagnostic tests (this or methodologically tight ones) is that they just label a status, but do not describe a process by which to move on. They can also be faked, intentionally or with the subject not beign aware, since they base on the intellect to chose the answer to questions. A demand by the tester that the subject should act spontane, does not change that.
I do not comment the computer-version of the test shown on page one, since I am suspicious to computerised versions of originally material, real-worl tests, in principal. I imagine the day when you go to an automat like to a banking robot, and you enter your chip card and get asked for your colour preferences and the damn thing next tells you who you are (or what it thinks you should be like), and that you should do this or that. THX-1138, the movie, on my mind.