What's the issue?
Often, the research question needs to be approached from many diverse perspectives, which involves using different methods and data sources. The benefits of using multiple data sources depend on what they add to a particular piece of research. This could be related to diverse research contexts or to different information about the same subject.
The combination may assume several forms, depending on the importance given to a specific method in the overall research and on the development of the research process itself. In what we might call a "sequential model", you may begin with quantitative (e.g. survey) to "map" a subject and then pursue with qualitative (e.g. interview) to "get deeper" into some topics. Or you may start with the qualitative (e.g. observation, interviews) to explore a given subject and then turn to the quantitative. Alternatively, in what we might call a "concurrent model", you may follow both approaches simultaneously, either to explore in different ways the same aspects of your subject or to cross-validate (or "triangulate") information gathered through different methods (Lobe et al., 2007).
Multiple data sources may also confront us with different perspectives concerning the same subject. In some cases the only choice might be to combine sources in order to get all the information we need about our research object. In any case, defining the status of different data sources is mandatory in order to articulate properly all the information available and needed.
Unlike cases in which we deal with different methodologies, we may combine different sources of information within the same methodology, as in the case of using different questionnaires to address the same problem. In this situation one must be careful to distinguish between the criteria used in the various sources of information (e.g. how a particular variable is measured in different questionnaires). When using different samples (collected over different periods of time), or a sample obtained in several populations, one is also combining different sources of information.
Questions to consider
However, one shouldn't forget that comparing different sources (containing data gathered for different purposes) is not exactly the same as comparing information from a single data frame. In the first case we are considering secondary analysis; in the second case we are actually comparing data within the same (or an equivalent) dataset. This isn't only a problem of considering different sample designs, but also of being sure if (or to what extent) data are comparable and in what way this comparison may be carried out.
Asking the same questions of different individuals also confronts us with distinct perspectives in relation to what apparently is the same activity/practice/event. For example, when you ask parents about their children's activities and compare the answers with the children's own accounts, discrepancies are common.
Pitfalls to avoid
People often overlook the fact that existing data can be used. They make use of multiple sources without having a clear goal of why they do so. They underestimate the complexity of such studies (qualitative, quantitative, parents, children).
Derbyshire, P. (2005). Multiple methods in qualitative research with children: more insight or just more? Qualitative Research 5(4), 417-436.
Creswell, J. (2008). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches. Thousand Oaks: Sage
Creswell, J., Plano Clark, W. L. (2006). Designing and Conducting Mixed Methods Research. Thousand Oaks: Sage
Greene, J. (2007). Mixed Methods in Social Inquiry. Jossey-Bass: San Francisco.
Lobe, B., Livingstone, S., & Haddon, L. (Eds.). (2007). Researching Children's Experiences Online across Countries: EU Kids Online Network.