In broad terms, observations may be considered a fundamental data collection method used to gain an understanding of participants, behaviors, processes, or artifacts (Angrosino 2004, 2007; Bottorﬀ 2004). Observational research takes a variety of forms depending on the object of analysis, and the approach adopted is often discipline speciﬁc. In the social sciences, scholars may opt to conduct naturalistic observation, whereby researchers observe participants in their natural setting (Mehl and Robbins 2012). This type of observation situates the researcher as a neutral outside party to a given activity or interaction, though achieving absolute neutrality can be diﬃcult. Creswell (2013) refers to this role as a non-participant observer, and suggests that the researcher need not be physically present. In contrast, Kawulich (2005) describes participant observation in which the researcher adopts the dual role of observer and active participant in activities. Here, the researcher operates from within the social interaction or context, which, as DeWalt and DeWalt note, aims to “develop a holistic understanding of the phenomena under study” (2011: 110). Brief mention should also be made of the notion of self-observation, in which participants observe their own behavior and act as informants (Rodriguez and Ryave 2002). Likewise, observational studies in statistics and medicine attempt to determine
causal relationships of speciﬁc treatments or policies under non-experimental conditions (e.g., Cochran 1965; Rosenbaum 2002; Patterson and Morin 2012). This more unobtrusive approach is contrasted with experimental studies, in which researchers actively manipulate at least one, if not several, independent variables or treatments while controlling the experimental environment to rule out other causes (see Chapter 19, “Experimental research” in this volume). Scholars have also focused on observations themselves as an object of analysis,
and epistemological reﬂections on ﬁrst and second order observations can be made (Luhmann 1990). Such inquiry contrasts ﬁrst order observations, which posit activities as facts, with second order observations, which acknowledge the researcher’s inﬂuence on the observation process. As such, second order observations suggest a constructivist epistemology (von Glasersfeld 1995). This
position is in line with post-positivist recognition of the potential biases in research design.