A good search strategy is to break down your research question into the following components. This search strategy works particularly well and was developed in the health sciences - but it can be a useful framework for breaking down questions in a variety of fields.
P - Problem/Patient/Population
I - Intervention
C - Comparison
O - Outcome
Let's look at an example. My research question is: Does exercise reduce the symptoms of restless leg syndrome?
P - The problem here is Restless Leg Syndrome. The question is not specific enough to draw a population/patient. It's fairly obvious from the question that we are thinking about a human population, but we won't include that component in our preliminary search.
I - The intervention is Exercise. Does exercise make a difference?
C - In this case, the comparison is with those who do not exercise. Let's leave this component out though, as it can be difficult to specify a deficit comparison without mucking up the search.
O - The obvious outcome here is reduced symptoms. Let's keep this in mind but not include it in the preliminary search.
The good news is a decent PICO search strategy starts with the terminology outlined in P and I, giving us:
Restless Leg Syndrome AND Exercise
