There are aspects of Psychology that are scientific (the biological stuff), and parts that are not (the introspective stuff).
Trying to figure out why that person is doing what they are doing is not scientifically verifiable outside of correlations.
Realizing that schizophrenics have enlarged ventricles could maybe be argued as scientific (although it has little actual useful value, aside from recognizing a correlation).
However, learning the way the brain works and recognizing that various parts of the brain deal with specific information is scientific. Using MRIs and whatnot to determine brain activity while performing tasks is observable and measurable. Furthermore, subjects with brain damage in particular areas support our hypothesis.
I remember an interesting case study where a guy was shown a rose. He was asked to describe what he saw. His explanation was a convoluted red shape, with a green apparatus that it seemed to be perched upon. However, he could not make the recognition that it was a rose. He was then asked to smell it, and he immediately recognized it as rose. (I believe this patient was an epileptic, and had his corpus callosum severed to stop the siesures. Basically the left side of the brain and the right side of the brain could not communicate).