There are many different approaches to conduct scientific inquiry, which all are legitimate. Research can be exploratory or confirmatory, experimental, causal or descriptive, seeking to be objective or seeking to be subjective, and apply qualitative, quantitative or mixed method approach. There are many differences how about these approaches - and scholars using them - understand “the correct” way of conducting scientific exploration. For example, different approaches emphasize the role of theorethic work differently: some think that theory should be first formulated into proper hypothesis, others see that the data should drive the analysis. Similarly, scholars might have (strong) opinions why an approach (which they use) is “the best one” to understand the complexities of the phenomena under investigations. Some even fight what is the “right” way to use their approach. This was a long introduction to say that this book tries not to advocate for a particular way of doing research, but embrace openess on these questions. Flexibility in these may even be a critical component for imagination: allowing interesting and relevant research to emerge.
No matter how the research takes place, empirical work focuses around gathering, analysing and interprenting data-broadly understood. Figure 11.1 shows the simplified version of a research process, but different elaborations of this process are commonly used in data science (Fayyad et al., 1996), political science, sociology, media and communication research, ethnography.add references Each discipline, subdiscipline or research approach may have their own terms which they use to describe their process, and different levels of depth. For example, data scientists can focus on various stages of data processing before it is ready for analysis through machine learning approaches (Fayyad et al., 1996), while grounded theory approaches provide a detailed workflow for conducting the data analysis, moving from open coding to axial coding and towards new theory development.
Programming and computational methods can help on any of these three stages of research. Skills developed when examining the elections data (in Chapters 2, and 8) provided many tools for data management and ensuring it is ready for analysis. For example, we have learned how to filter data based on criteria and reformat it to suite our needs. Beyond these, computational tools can be used to collect digital data, such as content in online services or data where a smartphone has been throughout the day. Regarding the two next stages, the four method families provide powerful tools to analyse data and can even help in interpretation. Programming is a powerful tool for any research question if the question requires working on a larger set of data. We will now illustrate some aspects which require further consideration when working on computational social sciences.