Psychology aspires to be a serious science, and the majority of academic and research psychologists use scientific methods to study human behaviour. However, psychology doesn't have one, dominant paradigm in the way hard sciences like physics or physiology do, and thus psychologists adoring to different theoretical systems might use different types of methods, some of them at variance with a positivist model of science.
Observation is a very important method and often-times the best way to explore a new subject. The subjects might be aware or not that they are being observed, and the researcher may take part or not in the activities (the former is called participant observation). Observation is often accompanied by note taking, audio or video recording. Data gathered during observation is then analyzed and often used to formulate more specific hypotheses for further research.
Clinical psychologists, just like medical doctors, commonly use personal interviews to assess their clients. Structured or semi-structured interviews are also used in academic research. Some psychological studies use surveys to interview a larger number of people, asking them a series of questions with closed answers.
Psychologists have developed a huge number of various tests and questionnaires designed to measure and assess various variables in a standardised way. This allows the researchers to compare the scores of an individual or a specific group to typical scores in a given population. Tests and questionnaires exist to measure intelligence, creativity, anxiety, extroversion, depression, dogmatism, preferred learning styles, temperamental factors, sexual identity and many more aspects of human personality, behavior and preferences.
Experiments are excellent for testing specific hypotheses about possible causes or factors affecting behavior. Experimental testing of the hypotheses is rightly considered the highest standard in scientific enquiry and good experiments give insight into behavior and allow for formulation of larger theories.
In a typical experimental study, scientists put groups of human subject into different experimental conditions and measure their reactions. This measurement may use psychometric tests, pen-and-paper questionnaires, observation or even using physiological measure like blood hormone levels or heart rate. The researchers then compare the differences between the groups, usually using statistical methods.
Many topics don't yield themselves to experimental studies, either because it would be unethical, take too long or be financially or practical impossible. In such cases, psychologists conduct correlation studies, which analyze past influences and differences between people in the light of a specific theory, testing predictions as to a character and direction of differences.
Popular among clinical psychologists, case studies analyze individual narratives, personal experiences or groups of people (families, teams, organizations) in order to achieve a deep insight into their particular dynamics.
Regardless of the specific methodology or technique used, all good research has roots in theory. Open-ended approach has its place in initial exploration, but there are always unspoken assumptions. In reality, nobody approaches human behavior subject with a completely blank canvas, least of all a professional researcher. Good study design is essential for meaningful and valid research.
Resources for further reading:
How Do Psychologists Study Behavior? at http://www.essortment.com/psychologists-study-behavior-35999.html
AllPsych: Introduction to Psychology and Research Methods at: http://allpsych.com/psychology101/intro.html
Cornell University: Social Research Methods at: http://www.socialresearchmethods.net/