Social scientists use particular methods to gather qualitative evidence, from observation to interview, but they also use autobiographical accounts, journalism, and other documentary material to flesh out and add meaning to statistics. As with reading numbers, reading textual evidence requires us to practice, to set time aside to learn how to do it, and to understand the conventions of writing which operate in the different forms of writing we encounter. One of the main problems with reading textual evidence, though, is that, unlike the relationship most of us have with numbers where we may use them at a pretty basic level, most of us are, if anything, over-familiar with words. When we want to understand their value as social science evidence we need to forget how familiar we are with first person accounts and everyday speech - for example, in newspapers, magazines, and books - and learn a different approach to them.
Social scientists use observation, interviews and even print journalism as evidence for the claims they make. They may collect evidence through questionnaires with pre-set questions and by open-ended interviews which allow respondents to speak for themselves. They may observe social relations explicitly as social scientists or may participate themselves in a particular community to gain 'inside' information.
Social scientists also draw on print journalism on occasion and may use the same sources, for example official statistics, and the work of other social scientists to support their claims. We need to remember, though, that journalists do not need to present the same rigorous referencing and support for their claims as social scientists are required to do. Most importantly, newspaper and magazine articles are written under commercial pressures; for example they must help to sell the newspaper by being deliberately provocative, or by reflecting the dominant views of its readers.
According to this passage, what do social scientists use written sources to do?