No one questions the value of observation as a method to collect data. The whole field of anthropology would likely cease to exist if observation were to be banned! For those of us who let video cameras do the observation for us, we can easily recite the advantages of video-based methods over direct observation: video records more information than could otherwise be captured (because of the density of information on video), and information captured on video is retained (permanence). Because of these first two, a third advantage arises: video can be reviewed many times over, to make sure that all relevant data are extracted.
So it might seem counter-intuitive to suggest that direct observation be included as a first step in a video-based method such as video reflexive ethnography. But I hope to convince you of the value of starting a video project with observation, so please read on…
Observation is critical for several reasons. In general, there is the need to address logistical issues ahead of time, so that when video recording begins, disruptions are minimized. Logistical issues include those directly related to the camera such as set-up, lighting and audio, and also include issues more related to the research such as figuring out when to obtain informed consent or understanding workflow patterns. Participants in video-based research projects often forget after a very short period (minutes) that there is a camera capturing their words and gestures. However, it would be difficult for participants to forget the camera if the camera was constantly moving or getting in the way of what it was the participant was trying to do. By observing first, a camera can be set up in a specific location that will be as unobtrusive as possible and minimize disruptions. While in some studies a person holds a camera and records the action, in other studies cameras are set up on poles or stands and set to record, then the person responsible for the camera leaves the room. This second approach is a great strategy for maintaining confidentiality, when recording in very small spaces, or when it would not be feasible to bring the camera any closer to the action (e.g., recording a surgical procedure where sterility must be maintained). Conducting observation first helps to determine where the camera should be placed.
Lighting and audio quality can make or break a video. If the camera lens is facing a window with light streaming in, the resulting image may be unusable because of overexposure. Similarly filming in a dark room poses its own unique set of challenges. Technological advances in video software can sometimes mitigate the effects of over or underexposure but by conducting observation first in the setting where filming will be done (and during approximately the same time of day) can provide crucial information about where to place the camera. For most video-based research, words and utterances are important data elements, along with the visual. Having adequate audio quality is another important consideration that can be facilitated by observation. For example, in my first video reflexive ethnography study, we followed physician teams around a busy academic medical center as they made their morning rounds, capturing conversations between physicians and the nurses who provided care for their patients. During our observation sessions we learned that the hospital corridors were noisy but the noise level varied in unpredictable ways, depending on factors such as who was walking by, how loudly they were talking, and the size and engagement of the rounding team. We realized that physicians would frequently continue conversations with nurses while walking into patients’ rooms, which meant that their backs would turn away from us (and the camera) as they talked. Without this information, the audio quality of our videos would have suffered. We supplemented the hand-held camera we were using with an omni-directional microphone attached to the camera (to pick up the voices of those who were talking without us having to move the camera to anyone’s face), and also with a lavalier microphone that was worn by the attending physician of the team we were following that day. Without having conducted observation first, we would not have realized the importance of the extra equipment to boost audio gain.
We’ve already blogged about getting IRB approval (see September 2019) but figuring out where and when in the research process to obtain informed consent can also be facilitated by first conducting observation. The goal is to minimize interruptions from the research team once filming has commenced, and obviously the process of obtaining informed consent comes before filming starts. But depending on factors such as the topic of study, the number of participants, and the location where filming will occur, there may be more than one approach to obtain informed consent, and the optimal process will become apparent through observation. Related to the informed consent process is the need to learn about participants’ workflow processes. For example, in our current VRE in oncology study we have identified a private consultation room in the busy oncology clinic where we will be able to consent patients prior to the study. We need to observe clinic and pharmacist workflow first though to learn what point in the patient’s visit to the clinic will be least disruptive for us to obtain informed consent. Talking to patients about the study, answering their questions, and obtaining informed consent is a time-consuming process and should not be rushed. On the other hand, having pharmacists sit idle, waiting to start their education session with a patient while we obtain informed consent would be disruptive to workflow, and possibly disrupt the clinic schedule for the day.
So how many observation sessions are enough? While there is no hard and fast rule about the number of observations to conduct, we (and others) have found that three observation sessions can provide enough information to address most logistical challenges and variations in the context of what will be filmed. Of course, something unique and unexpected can always crop up (we’re talking about research in healthcare after all) but the advantages of conducting enough observations ahead of time have to be balanced against the costs and resources expended.
In summary, video-based methods have replaced observation for data collection purposes in many instances, and will continue to grow in popularity as new applications emerge and current applications are refined. However, ignoring the power of observation as a primer to video-based methods carries risks to the reliability, validity, and credibility of findings.