Good concept (anybody need a dissertation topic?). I don't remember seeing anything written about this but I've discussed it in workshops a few times. It makes sense theoretically and I've found it useful clinically a few times.

One major problem with projectives is the reliability and validity (or lack thereof) of interpretive systems. For this approach to projectives to work well, someone would need to come up with a systematic approach to scoring and interpreting the client's responses so that different clinicians show an adequate level if interrater reliability and then would need to demonstrate an adequate level of validity. I bet it can be done, but it would not be a small project.

If this can be done, another important question will be whether time spent administering and interpreting the TAT is productive enough to justify spending session time in that way. My bet is that there are more efficient ways to get the same information. It probably will turn out that it is more productive to spend session time pin-pointing automatic thoughts that occur in problem situations rather than spending the time interpreting responses to the TAT.

However, it is possible that this TAT methodology could be quite useful in research. CT researchers really need valid ways to assess dysfunctional beliefs, schemas, interpersonal strategies, etc. without just relying on self-report. Structured methods for interpreting responses to TAT-like cards, sentence completion tests, etc. would have a lot of potential.