Well psychologists follow the DSM as well, at least in the US. I find that diagnosis is often disagreed upon among mental health professionals. I dont know if there is much empirical evidence to support the notion that psychologists and psychiatrists disagree at a differentially greater rate than what is normal. At least for pure psych diagnosis. This may just happen to be your personal experience thus far. That said, I think psychologists may be in a better postion to rule out malingering or symptom exaggeration because there are many psychmetrically sound instruments (and validity scales within scales) that can asssit with this process that psychiatrists are not trained on. I also think psychlogists may be better (or more clinically cautious rather) about assigning personality disorders, as we typically spend more time with patients and have a number of psychometrically sound instruments (ie., MCMI-III) that assist with teasing these issues out from a complex clinical picture.
I also think "clinical judegment" may be used somwewhat differentially between the 2 proffesions at times. For example, if I am doing a SCID and I ask the first critria for MDE in Module A ("have you been feeling down or deprssed everyday most of the day this past month, etc) and the person anwers yes, I dont take that at face value and jsut move on. I contuinue to digg and ask questions until I am satisfied that they truely met that critria as it is meant clinically. In otherwords, I make the judegment, not the patient. I dont generally take what they say at face value right off the bat and put a chech mark next to it an move on. I am sure most psychiatrists take the time do do this as well, but I have seen many that do not.
Lastly, neuropsychologists are often called upon for differential diagnosis of complex neurologic and cognitive disorders such as AD and other dementias because many neuropsychological tests can establish what parts of the brain are being affected. Thus suggesting the undelying neuropathology, and giving hints about what disease process is taking palce. Obviously, this presents a benefit over something like the MMSE, since it is a global meaure of impairment (kinda like the Glascow Coma Scale) and has zero specificity for any one condition. In other words, the MMSE can tell you that the pt is having a problem, but cant tell you why, or what brain systems are effected. Hence, it doesnt tell you anything about ther disease process. This is why many neurologists will refer to a clinical neuropsychologist for more indepth testing that will help pinpoint the specific systems that are imapired in order to get a differential diagnosis.
UPDATE: I think OLLIE makes a great point. Spending time delineating the anxiety from the depression is not very fruitful if you are simply gonna provide a short supportive check up and write a prescription for zoloft or Wellbutrin. However,for psychotherapy, the inteventions for the 2 issues will be quite different, and the 2 might have very differrent triggers and origins. Hence why psychogists (since they are primarily the ones doing long therapy these days) will want to spend more effort teasing apart these issues.