Harvard Library licenses access to hundreds of databases. Although access usually requires logging into each database one by one, there are several noteworthy databases for social science research that enable access to dozens of datasets through a single interface. These tend to be more "plug and play" and are more efficient for downloading large amounts of data. Some common characteristics:
Data repositories enable social science researchers to securely store and share their datasets. These repositories are good places to identify and access data that have been previously collected and used in academic research papers and projects. While some repositories are discipline-specific, others are broad in subject scope; some may be institution-specific while others are open access.
For more data repositories, visit the Research Data Management @Harvard site.
In addition to searching licensed databases and data repositories, a literature search is a useful technique for identifying datasets and research methodologies in your discipline. These "aggregators" enable discovery of data cited in academic research output such as peer-reviewed journal articles, working papers, conference proceedings, etc.
A critical aspect of scholarly research is creating good documentation about your data and your data management practices. This is especially important for reproducibility of results, data integrity, and is often required by publishers and funding sources (and also saves you time and aggravation along the way!). A few resources to get you started: