In my oppinion the answer is very much dependent on the method you apply. Or rather it highly effects the way you gather data about your client. Working mostly in a psychodynamic approach I would suggest that transference - countertransference issues are very informative data and you should apply specific "techniques" to strengthen these processes. E.g. not only ask about life history, but address your relation with the client. "How did you feel about being with me?" "What kind of fantasies did you have about who you will meet as a therapist?". To address this issue is also important because it is a learning process. If you fail to address it in the beginning, it will be very hard to talk about client's negative feelings toward therapist. On the other hand, a good diagnosis or formulation is inevitable to plan and carry out treatment. We should find the tests and methods that best fit our therapeutic thinking or method.

I do not focus much on assessments. I instead stay with the client's needs. I ask them what (all) they would want from therapy. I carry out a thorough assessment interview collaboratively so the clients do not feel they are under the scanner or the are being 'analysed''. I find the information collected through an interview much more useful as well as richer to use in therapy.

I must also add that it does feel a bit anxiety-provoking in the beginning to decide about whether or not to conduct any standardised tests before beginning therapy. Being in supervision with a competent supervisor often helps.

When I trained as a clinical psychologist some thirty years ago, we were required to write case studies that contained an assessment and a formulation an intervention and some sort of outcome. Of course the problem is that as soon as you start to assess you are intervening and as you intervene you will gain new information that changes your assessment and your formulation. My sense then and now is of a continuing iterative process, where intervention information and formulation continue to evolve sometimes up to the point of termination where the client and the therapist are still learning.

All the same, there are times when a good assessment is very helpful. This is particularly the case where a number of competent therapists have tried and failed to help the client to make progress. An analogy is the instruction manual usually not necessary but sometimes vital to avoid making mistakes.

When an assessment is necessary I follow a protocol borrowed from Lester Luborsky and supplemented by some ideas from the late Mary Burton. I am not sure that the precise format matters but I am interested in where people come from, particularly early attachment, what they have made of their lives, the history of their problems and of their attempts to solve them. I want from the assessment to begin to construct narratives and hypotheses about their lives and their current issues.

When I was younger, I took the view that we should only find out enough about a person to enable us to make an assessment that would improve our decision making and thus the course of therapy. I still feel that it is good not to be over intrusive and yet perhaps being reticent in the questions one asks can be a bit collusive.

As a researcher and a practitioner researcher, I have always been troubled by the distinction made between assessment and therapy. Assessments are exercises in meaning making and are mutative. They are an opportunity for the client to express emotion and be understood, as such I would want them to be measured as part of the change process.

I generally deplore the notion of "assessing for suitability for therapy".

I suscribe with many of the ideas posed by Jeremy. The idea that the assessment process and actual intervention are two distinct moments is illusory, as our mere presence and questions can provoke profund changes. I also suscribe the idea that the therapeutic process is an open ended process of exploration into the patient's inner world, so the assessment should be about how much do I need to know (enough and sufficient) to start working with the patient. I also agreed with the idea that assessment should be to see if we can help and in which way, and not about if the patient suits our particular technique or needs.

In my case it can take 2 session (I try to have at least two separate encounters with a patient) or more. I don't pre-establish a duration, but it shouldn't go on forever. If I'm not concluding, I try to understand why it's happening, by myself or with the aid of others (supervision, peers). I think we all do assessment, there may be differences in terms of the framework (diagnose, inner dynamic, relational, problem focus, experience focus, etc), but at the start of our work we all try to understand what's the problem before actually deciding to do something.

I also think, as András mentions, that considering transference - countertransference dynamic is essential part of psychotherapeutic work, but I would differentiate between observing and managing transference - countertransference and overtly exploring and interpreting it. Some type of patients react with a lot of anxiety when confronted with some aspects of reality, e.g. the emotional bond with others, that they have been denying or warding off, so confronting them too soon is actually like showing them the door out.