Setting the primary direction of research starts with the choice of the topic. Although interesting and important research have been conducted in the most esoteric of areas, research students should choose their topic with care. The topic should be of interest to the researcher and his supervisor. This will minimize the researcher's boredom and maximize support from the supervisor. Choosing a topic area of some relevance to the discipline or to the society is always a good idea. There are fashions in science and working on a popular subject can boost further career, but also puts competitive pressure on. Examples of topic areas in social sciences include: correlates of intelligence, changes in sexual morality, media and the beauty ideals, geography of urban poverty and affluence and women in management.
Having chosen a topic area, a budding researcher needs to narrow the direction of research down, leading eventually to the formulation of the research problem. The research problem is, in simple terms, the question that the study will answer: What is the relationship between the personality traits and intelligence? The hypothesis is the prediction about the nature of the problem that the research will test: Extroversion and introversion have no effect on IQ, but neuroticism will have a negative connection to IQ, particularly non-verbal scales.
In hard sciences, there is usually one dominant paradigm in any given discipline; in social sciences choosing one's theoretical perspective is normally a necessary part of focusing the research.
Reading, or reviewing existing literature, is absolutely crucial to the setting of research direction. The best strategy is probably to start with general material-- usually books--and to proceed to more specific texts--usually papers in peer-reviewed journals. This process should accomplish three goals. Firstly, it will show what concepts and terminology are used in the area. Secondly, it will familiarize the researcher with the main theories and the current state of knowledge. Thirdly, it will show the popular research themes and pinpoint gaps in the knowledge. Many studies conclude wit recommendations for further research and such recommendations will often provide inspiration sufficient to set direction of your studies. In some cases, a completely new approach will seem appropriate and might be the most fruitful route: just because nobody did something before doesn't mean it can't or shouldn't be done. Such cases, need especially careful literature review and a strong input from the supervisor.
A good supervisor will be an invaluable help at the literature review stage and should point the young scientist to important, interesting and stimulating materials. But the supervisor can do much more than recommend reading: she can be a sounding board and help the young researcher shape his initial ideas and direct them towards innovative, interesting but doable forms. The supervisor will also have a much broader knowledge of the subject's literature and thus if the student researcher comes up with an idea or a problem that has been already researched, but one he didn't come across, the supervisor will be able to judge whether further exploration makes sense. Often, there will be other researchers working with the same supervisor. Using them to bounce the ideas off and share the reading experience will also help with problem and hypothesis formulation.