The article deals with the question of linking two spheres: research practice (psychology as empirical science) and professional practice1 (psychology as a scientifically and ethically meaningful practical action: diagnostic and therapeutic). The author assumes that the most important task of psychologists is to create testable (in Popper’s sense) empirical theories. Only based on these can professional practice be built (here: clinical – diagnostic and therapeutic). The effectiveness of clinicians’ professional actions in the sphere of social practice is a derivative of the method of practical action: diagnostic and therapeutic, built on this knowledge. The author formulates a strong thesis that beyond the context of proven empirical theory, there is no sensible, and yet ethical, professional practice. The article consists of two parts. The first part deals with the methodological aspects of the relationship: empirical theory – testing of the theory – professional practice. It also applies to evidence-based assessment (EBA) and evidence-based practice in psychology (EBPP). The clinician uses a variety of data in his/her (scientific and practical) work, which raises the problem of how to integrate these data. This is the second part of the article. The author distinguishes four levels of data integration: (1) construction of variables and the building of hypothetical relationships between them; (2) operationalisation of variables, i.e. the transmission of variables from level I of the empirical sense; (3) quantitative interpretation of empirical research – here the interpretative framework is the psychological test theory (or another tool used in the operationalisation procedure); (4) qualitative (clinical) interpretation developed on level III data – the psychological empirical theory here provides the interpretative framework. At each level, we are dealing with theories. The empirical data that emerge are indirectly brought forward and justified by these theories. These levels are somewhat dependent on each other. In other words, we are dealing with integration within each level and between levels.