Participant and non-participant observation

There are two distinct approaches to observation: participant and non-participant (or naturalistic). This chapter focuses on non-participant observation, as it’s more appropriate for the type of research questions that designers are likely to ask. It entails observing people’s behaviour in real-world settings, fly-on-the-wall style, and was integral to the landmark studies mentioned above. As ‘non-participant observation’ implies, researchers unobtrusively observe people in their everyday lives, doing whatever they normally do, uninfluenced by the researcher’s presence. This can take either a structured approach to gather mainly quantitative data or an unstructured approach to gather qualitative data (see Figure 3.1).

Structured observation

Structured observations tend to focus on human activity that can be systematically mapped, counted and measured. This might then be used to assess pedestrian volumes or common routes, or gain insight into specific topics like the types of activities in a space, how long people stay there, or demographic trends, perhaps. Information from structured observation can also inform decisions on aspects of a design or site layout. Do features within a space create physical barriers or encourage interaction? Do the ways that different groups use the same space conflict?

Figure 3.1 Types of observation

Some observations can be undertaken perfectly effectively using the low-cost no-frills techniques favoured by Gehl: human observers, in-situ, with pen and paper. There’s the added benefit of having people on the ground who are able to help interpret anything unclear. Complex sites or large-scale projects are likely to require more sophisticated survey technology. GPS, CCTV and the like allow the collection of much greater volumes of data over longer time periods, with counting and mapping automated or done remotely. Both hi-tech and low-tech strategies should produce the same result, however, namely reliable data that provides a foundation for evidence-based decision-making.

Unstructured observation

Unstructured observation offers real benefits in the exploratory phase at the start of a project. Effective observers can provide valuable insight into local social-spatial dynamics to help the team understand the realities of the site context and to consider options for the site in this light. At the same time, it’s also an excellent way to begin developing ideas about approaches to engaging with the local population, keeping the location, the community and the new development at the forefront. It gives observers the freedom to note anything that might pertain to the research questions, and allows them to form their own impressions of a site’s significant aspects rather than starting with a list of things on which to focus. Of course, designers don’t need to be told to look round a site beforehand and scope it out for themselves, so isn’t that an unstructured observation? The difference is methodological rigour. Observation is purposeful and focused, guided by defined research objectives, yielding copious notes and other records, which are methodically analysed in depth. So it’s much more than just having a look round!

Preliminary observation

If possible, members of the design/observation team should undertake a preliminary observation visit to inform the observation proper. This will clarify its aims and parameters, defining the questions it will be answering, and deciding how to capture the information those questions require. This should include quantitative work, such as:

• mapping and photographing the site
• identifying the main routes and active areas
• selecting observation locations
• setting the boundaries of the observation site.

There is also qualitative work to do towards understanding the socio-spatial context, for example noting visual and sensory impressions, sketching, watching interactions and behaviour, and generally starting to think like a researcher. Having clarified the observation’s aims and scope, the team should then be able to finalise the strategy and decide roles, timing and duration of observation periods, and identify resources required. Figure 3.2 shows a suggested process for the whole exercise.

Remember that the logistics of data gathering influence the information yielded. Factors like time constraints, observers’ locations and decisions on what to count or map often have implications, so consider what these could be. Observers will need an outline map of the study area showing the main paths and features within the space, and a reference sheet, tested out first whenever possible, setting out how everything should be recorded. Run a briefing session beforehand to ensure that observers will work consistently. It’s sensible at this point to discuss situations that could potentially arise for them, such as responding to questions about their activities, dealing with people who object to being watched or recorded, and personal safety concerns. Unlikely eventualities, but it’s better for observers to be prepared.

Bias issues

At the data-gathering stage, an unstructured observation can be influenced by observers’ preconceptions or misinterpretation. A structured observation can also yield flawed information by not including significant aspects of a site, ignoring types of users or overlooking types of use. Later, at the analysis stage, observers can be prone to jumping-to-conclusions types of errors, only seeing what they expected, or misattributing causes or meanings to what they’ve observed. These kinds of errors tend to occur when unquestioned assumptions or selective attention impacts on the data gathering or analysis, and are inevitable to some degree. Ascribing meaning or causality where meaning and causality cannot be known are constant temptations in an uncertain world, but have no place in methodologically robust research.

So how can these biases be tackled? A good level of self-awareness, integrity and attention to detail always make a successful observation possible. It might help to employ Popper’s method of falsification or refutability (Popper, 1959), whereby researchers play their own devil’s advocate and look for explanations to disprove their theories rather than evidence to support them. Another solution is to ask others to review the material and see if they arrive at similar conclusions. I would also endorse Gehl and Svarre’s advice on addressing an historical imbalance; disabled people, elderly people, women and children have been routinely overlooked throughout decades of urban planning and placemaking, so pay extra attention to these groups to redress the deficit.

Counting

Counting generates quantitative data on human activity in a space, useful for finding out things such as the times of main usage, who uses the space, whether there’s a broad or narrow range of users, the features that people do and don’t use, and the types of events and activity that occur there. These kinds of statistics can be a worthwhile addition to an evidence base informing decisions about a project, and provide baseline data to assess a scheme’s impact post-completion. Although counting can reveal patterns of activity, it doesn’t offer explanations. Nor can observers always identify with certainty what people are doing just by watching them. Interpretations are therefore usually provisional.

Jan Gehl (2011) characterises two types of activity in public space worth recording in structured or unstructured observation:

1. Necessary. Purposeful and goal-orientated activities such as essential shopping, and work-related journeys or tasks.
2. Optional. More leisurely or informal activities such as strolling, eating, playing, sitting, reading or chatting.

Gehl recommends noting the locations where optional activity occurs, especially the places where people choose to sit or stay – whether on the edges of a space or in the middle, in secluded areas or out in the open, for example – as signifying features that people value, such as sun, shade, sight lines and proximity to other activity. (Whyte, 1980, suggests the number of women sitting alone in a public space can indicate the sense of safety it offers, which may or may not be worth keeping in mind.) The ratio of necessary-to-optional activities show whether its character is primarily functional, enabling mainly necessary activity or whether its features encourage optional activities.

Activity can be measured by two main counting methods. The nature of the space, the topics of interest and the research questions will largely decide which is appropriate (it may be both).

1. Time-based counts note how many times particular activities or events occur in a given timeframe
2. Interval-based counts note everything that’s happening at given times during the day, for example, every 30 minutes.

A hand-held click-counter can be useful in busy spaces with large numbers of people; otherwise a pad and pen, or a pre-printed table or checklist to mark up should suffice. It’s best to undertake counts at different times of day and in different seasonal conditions, which means it can take several weeks or longer to gather sufficient data. Whatever the approach, it’s essential that all observers work consistently, counting and recording in the same ways.

Mapping

Mapping generates quantitative data on movement paths through a space, such as pedestrian flows and routes. It provides information on stationary behaviour too, showing areas where people linger and those they avoid. It can highlight problems in accessing the space, such as a lack of road crossings or clear paths, or hindrances within the space itself, such as steps or inconveniently placed fixtures. Again, undertake mapping exercises at different times of day and in different conditions, and if possible over several weeks. As with counting, mapping movement in and around a space can provide good baseline evidence to inform decision-making, and assess the completed scheme.

Data can be collected by one or both of these approaches:

1. Time-based mapping, when observers trace the routes taken through or around a space, recording individuals’ approximate paths on a map. Although inexact, it can provide a clear enough picture of primary and secondary routes, and paths less travelled. In spaces with large numbers of people, some systematic random sampling may be necessary: tracking every Nth person who enters, perhaps.
2. Interval-based mapping, when observers mark up a map at specific time intervals to show the locations of different groups of users and types of activities at that moment, using an agreed notation. These produce an aggregated picture over the whole time period when merged, so this method provides information on users, activities and both datasets combined.

Gehl and Svarre also suggest measuring walking speed in pedestrianised settings as an indicator of the quality of a space, on the basis that a slower pace indicates a more enjoyable environment. Calculate this by finding an observable stretch of walkable space and measuring out a section of approximately 100 metres, then measuring with a stopwatch the time it takes randomly selected pedestrians to walk from end to end. They also recommend test walks, where observers walk key routes themselves, noting positive and negative features along the way. This can highlight some considerable differences in walkability between how a route looks on a plan and the reality on the ground. For example, a test walk for a study they undertook in Sydney produced the gloomy finding that 52% of the total walking time on one route was spent waiting at pedestrian crossings.

Photography and video

Video and still photographs can record structured or unstructured observations to provide an accessible visual documentation of a space and the nature of activity within it. This can contextualise other quantitative information, or provide a data source itself. For example, a series of captures of a particular activity can reveal new information about that activity when images are examined together afterwards. Images might be recorded to a brief, to address specific research questions or to capture particular types of events, or could adopt a less structured approach to depict a range of aspects or qualities of a space. The AEIOU mnemonic (Martin and Hanington, 2012) is a useful reminder for observers to document:

• Activities
• Environments
• Interactions
• Objects
• Users

Shoot more images or footage than is strictly needed, as aspects can be noticed in review that were missed on site, so it’s sometimes helpful to have extra material to examine. Aesthetic value is of no importance here; images just need to be clear, well lit and properly exposed. In a small space, a smartphone may be suitable for recording video or stills if the image quality is adequate, and will attract less attention (and therefore potentially influence people’s behaviour) than a video or camera.

Photography or video recording for observation exercises requires discretion and sensitivity; even in a public space, there are ethical issues to consider and limits to what it’s acceptable to shoot. It’s legal to photograph people or anything else from a public space without consent, but this can still be an invasion of personal privacy. If in doubt, let people know that you’re documenting the general environment, not them personally. Whyte recommends not recording people in ignominious situations or in the midst of personal difficulties, or engaged in any kind of illicit activity, even if these do take place in public.

Simulation and experiential monitoring

It is now possible to record individuals’ cognitive and emotional responses as well as their patterns of movement. Thanks to rapidly evolving technology utilising various types of sensor linked to a phone app or portable device that can monitor brainwave patterns, heartbeat, eye movements and breathing, people’s responses to different real-world environments can be measured, showing the levels of interest, pleasure or stress that their surroundings evoke. Similarly, experimental urban design research creates immersive virtual environments in laboratory settings to record people’s physiological reactions to simulated urban spaces and experiences. An unprecedented depth of insight into wayfinding behaviour, cognition and the effects on mood that different street typologies and environments create is now available. Yet the director of one such research centre – Colin Ellard at the Urban Realities Laboratory based at the University of Waterloo in Canada – maintains that ‘Although there is no doubting the power of a virtual reality simulation to unearth relationships between the organisation of the built environment and the operation of our minds … there is no substitute for experimentation at street level’ (Ellard, 2014). The knowledge generated from this field of research provides designers with exciting opportunities to create people-centred environments offering maximum enjoyment and interest, and minimum stress and anxiety, taking much of the guesswork out of whether the public will embrace or avoid a new space. Nonetheless, observation of real world street life, with its traffic, noise, smells, crowds, random interactions and unexpected occurrences remains indispensable.