Document analysis for what? Document analysis is undertaken in three distinct circumstances.
First, when the documents themselves are the focus of the research. This occurs, for example, in media studies where newspaper coverage of an event is scrutinised, where the sexist bias in advertisements is analysed or the ideological content of films is deconstructed. Similarly, a specific document, such as a policy document, may be the subject of scrutiny.
Second, when the documents constitute the main evidence in an enquiry or additional evidence to support other forms of research such as observational or survey data. Historical research primarily uses document analysis and it is regarded as a key skill in historical interpretation. Document analysis is also the most frequently used data collection measure among political scientists.
Document analysis in this sense is distinct from literature review, which essentially provides the context in which further research takes place rather than adds data to the research enquiry. A literature review context may indicate gaps in the research area, identify theoretical perspectives adopted or provide outcomes against which the new research is compared but it does not add new data that is part of the analysis.
Third, document analysis is used when the research enquiry involves undertaking a meta-analysis of all (or a specified subset) of the available published (and grey) literature on a subject. For example, Lee Harvey (2009) undertook a wide-ranging analysis of the existing research on Foundation degrees in the United Kingdom that encompassed all 317 available studies on the new qualification. Again this differs from a literature review in that it synthesises the material and presents the results as the research outcomes, rather than simply contextualising the research project, which is the point of the literature review.
Using documents for research has advantages and disadvantages. Documents exist independently of the research enquiry and provide evidence at minimum cost to the research team (that is, the researcher does not have to collect first-hand data). Documents provide access to data that would be difficult or very time consuming to gather via first-hand research methods, such as interviewing or observation. In the case of historical analysis, documents (including eye witness testimonies or oral accounts) are likely to be the only source of data.
However, documents can be difficult to access and often the researcher is unaware which documents exist that are pertinent to the enquiry, these may be discovered in the course of an archive search. For example, Harvey (1987) undertook a study of the Chicago School of Sociology and discovered by chance that William Ogburn's journal, which he did not know existed and which had 21 year embargo on it, was made available the month in which he visited and worked in the Regenstein Library Special Collections. The journal proved an invaluable source for debunking some myths about the Chicago School.
There are also likely to be gaps in the documented record, either because documents have gone missing, never existed in the first place or are in scattered locations. Some documents may have been accidentally or deliberately destroyed if, for example, the content is embarrassing or controversial. Some documents may be classified and not available. Some documents may be inaccurate, misleading, falsified, redacted or even forgeries. For this reason, it is important to undertake an external analysis of documents used prior to undertaking an internal analysis of the content.