Abstract
Non-participative observation, a data-gathering tool where the researcher assumes the role of observer while minimising participation, has a relatively low profile in empirical political research. This article begins with some introductory comments on the status of non-participative observation in political research. It then sets out the particular attributes of non-participative observation, focusing on the practical application and utilisation of non-participative observation in empirical political research. I show that non-participative observation can make a valuable, if particular, contribution as a data-gathering tool in empirical political research.
Non-participative observation: the ‘poor’ relation
Too rarely do political researchers adopt the role of non-participative observer. Other approaches to data collection, such as interviews, surveys, content analysis and participant observation are more readily utilised. Non-participative observation (NPO) is, I contend, a neglected data-gathering tool in political research, but unjustifiably so, because it has particular attributes and utility. It does not deserve its current position as the ‘poor relation’ of political research methods.
NPO is often presented as only a marginally useful data collection tool, the last resort of the political researcher. For example, according to Peter Burnham and his colleagues (2004, p. 232) NPO is:
‘a category which describes situations where the researcher has failed to negotiate the access he or she would have liked, and so conducts the research at some distance from the community or group under investigation. … The role of non-participant observer is thus often imposed on social science researchers, because either they have failed to convince groups to allow them access, or because of the nature of the situation in which the research is being conducted’.
Although they do not entirely dismiss it as a research method, Burnham et al.'s statement demonstrates how NPO can be misrepresented. There is, additionally, a tendency in political research and general social science research texts to focus on other methods, relegating NPO to sometimes the briefest of mentions. 1 It is these propensities towards misrepresentation and neglect, rather than any fatal flaw in the approach itself, that lie at the root of NPO's ‘poor relation’ status in political research.
Empirical political researchers, with varying degrees of emphasis, are interested in describing, explaining, understanding and evaluating political ‘institutions, attitudes and behaviour, policy making decisions and networks’ (Blaikie, 2000, p. 72; Harrison, 2001, p. 2). Important in such endeavours is a ‘family’ of data-gathering tools, each with an established set of attributes and applications, that researchers can turn to when planning data collection strategies. Observation, including non-participative approaches, can contribute to political research and therefore ought to be recognised as a closer member of this family (Dargie, 1998, p. 66). In this article I intend to highlight the attributes of NPO with reference to specific examples of its (rare) practical application in political research projects. Briefly, these are Lyn Kathlene's (1994) research on ‘conversational dynamics’ in US state legislative committee hearings, Emma Crewe's (2005) anthropological study of the House of Lords and my own current research into democratic processes in Mossbank, part of a small Scottish town. 2
What is non-participative observation?
Watching and listening to what people do and how they behave in particular settings and situations is the foundation of all observational methods (Robson, 2002, p. 309). Observation can be discussed with reference to several ‘ideal-types’, usually presented as opposing pairs. 3 One of these dichotomies focuses on the extent to which the researcher participates in the setting under observation. NPO represents one extreme, and participant observation (PO) the other. In the former, the observer aims to observe while remaining aloof and uninvolved so as not to influence the setting under study (Bryman, 2004, p. 167).
This defining aspect of NPO means that the researcher can observe behaviour or processes as they would ‘naturally’ occur (Shaughnessy, Zechmeister and Zechmeister 2006, p. 108). Interviews, surveys, even PO cannot do this. Thus, a non-participative role can minimise ‘reactive effects’, changes in actors' behaviour due to the presence of the researcher (Bryman, 2004, p. 175). Data from interviews and surveys focus on what people say they have done or will do. In PO, the observer sets out to become involved in, or influence, the setting under investigation by striking up relationships with people in the setting (Burnham et al., 2004, p. 222). Participation can get in the way of observation because the researcher has to balance the two roles (Gomm, 2004, p. 229). For example, Crewe participates in the activities of the House of Lords to the extent of gaining employment in a number of departments. This means that not only does she have to carry on with observing, she has to, at various times, work for ‘Black Rod's department, selling pictures to peers’ and ‘the Crossbench Convener as a personal assistant’ (Crewe, 2005, p. 246). In practice, NPO differs from PO in the degree to which the observer interacts and participates in the setting under study – I return to the question of ‘degree’ later.
Attributes of non-participative observation
Data collection in political research, as much as in any other field, is a practical undertaking that is constantly adapted and refined. This section is divided into three parts, each concerned with an aspect of practical data collection. I first discuss practical reasons for using NPO, focusing on issues of access, commitment and time, as well as its advantages of non-interference. I then discuss some different methods of producing data through NPO. Finally, I set out the benefits – and some dangers – involved in taking a flexible approach to the use of this method.
Access, commitment and time
NPO can be used to collect data when they are otherwise hard to obtain. This is particularly the case where the researcher cannot gain access to the process under study, when participation is unethical or dangerous or when co-operation is difficult. Crewe, for example, can access almost all aspects of the House of Lords. However, she is not a peer, so she cannot access or participate in the parliamentary ceremonies. Her comments relating to the state opening of parliament are revealing on this score: ‘In 1999 I watched the procession to Westminster from the Clerk of the Parliaments’ office window' (Crewe, 2005, p. 210). In relation to the previous year's state opening, Crewe describes how she ‘sat in the Chamber itself, and watched from close quarters the climax of the whole event’ (Crewe, 2005, p. 210). She is not participating in the pageantry. Adopting the role of non-participant observer is Crewe's only option if she wants to observe these formal events.
Some processes or behaviour of interest to political researchers occur only intermittently. Periods of data collecting can be spread over time matching the cycles of the process of interest to the researcher. Under such circumstances, PO may be rejected as an inefficient use of the researcher's time and resources. The time and resources expended in accessing the setting and building relations with participants may outweigh the return in terms of raw data. The researcher may also impinge unnecessarily on the processes under study. However, NPO affords the researcher first-hand access to these processes without the preparation and commitment required of PO and with less likelihood of unnecessarily affecting the processes under study. For example, while preparing to observe the monthly meetings of the residents' committee in Mossbank, I rejected PO in favour of a non-participative approach. Adopting a non-participative role, I could access the meetings while avoiding the difficulties associated with joining and taking part in the business of the committee. Taking part in the decision-making processes central to the activities of the committee would have diverted my attention away from observing the behaviour of the individual members. Adopting NPO allowed me to concentrate on the task of data gathering.
Data production
Like many other data collection methods, such as interviews and questionnaires, NPO can produce data in the form of words or numbers. Whether the focus is on content, process or the behaviour of actors in a particular setting or situation, the non-participative observer is not constrained in how they wish to record raw data (Bell, 2005, pp. 188–189). When collecting data as numbers, the non-participative observer can adopt a ‘structured’ approach, counting observations using preprepared schedules.
Kathlene, while researching conversational dynamics in US state legislator committee debates, records her data as numbers using a schedule listing five forms of interruption: ‘smooth speaker switch’, ‘simple interrupt’, ‘overlap’, ‘silent interrupt’, ‘talkover’ and ‘butting-in’ (Kathlene, 1994, pp. 564–565). As a non-participative observer, Kathlene has adopted a role suitable for counting both the number of interruptions and the nature of the interruptions relating to individual speakers in the hearings.
In less structured approaches to NPO, the researcher records data as words (Bryman, 2004, p. 167). Following common practice, I relied on a set of prompts to guide my observations of the residents' association in Mossbank (Marshall and Rossman, 1999, p. 107). 4 Prompts act as reminders to the researcher as to what particular aspects of the process and setting are of interest. I attended the committee meetings with written prompts designed to focus my observing. These prompts include reminders to observe the layout of the meeting room, how the members arranged themselves in the room, how the members were dressed, their demeanour, who talked and who did not talk, non-verbal communication between members, networks, factions or divisions noticeable in meetings and who was engaged and disengaged.
NPO does not impose constraints on the form in which raw data can be recorded. It is therefore an option for researchers who use qualitative- (words and language) based research methods, for those who wish to use quantitative- (numbers and counting) based research methods and those who favour a combination of both.
Flexibility in non-participative observation
In practice, the role of non-participative observer offers the researcher considerable flexibility while preserving NPO's particular advantages as a data collection tool. Appreciating the flexibility of this role lies in viewing it, not as a rigid straitjacket, but as a ‘tendency’. Depending on the judgement of the researcher and the given circumstances, there is scope for a degree of participation while engaged in non-participative observation. The researcher may be confronted with developments, decisions and dilemmas that introduce or necessitate some form of participation. However, NPO, like other data-gathering tools, has inbuilt flexibility that can help the researcher cope with unexpected surprises or make the most of unforeseen opportunities encountered while collecting data.
The non-participative observer can minimise or rule out any risk of participation by remaining at a distance from the situation or actors under study. Audio and video recordings can be useful in this regard. Kathlene, for example, uses audio recordings as the basis of her data collection and analysis, and does not actually attend the meetings in person (Kathlene, 1994, p. 562). Covert observation, where the observer collects data without the knowledge of the subjects, is another possibility. However, practical and ethical issues limit its application in political research (Harrison, 2001, pp. 82–83).
Close contact with a setting is more likely to result in the non-participative observer tending towards some degree of participation. Acknowledging, but at the same time, controlling and minimising instances of participation preserves rather than undermines the non-participative role of the researcher. As Lisa Harrison (2001, p. 81) points out, a silent, unresponsive note-taker, too attached to the ideal of NPO, may actually, as a result of this behaviour, attract attention to themselves, creating unwelcome reactive effects and jeopardising trust and goodwill built up between the researcher and the actors in the social setting. Sometimes participation, even of the most minimal kind, is necessary. For example, the residents' committee that I observed would occasionally invite guests to their meetings. It was usual that introductions were made and I found that I would be included in these. The chairperson of the committee would introduce me to the guest: ‘Peter's here observing for his studies. Just to let you understand he's not part of the meeting’. In this particular instance, my response was as brief and unobtrusive as possible: I simply nodded to the guest.
If dealt with appropriately, moments tending towards participation do not cancel out the attributes of NPO. The tendency towards participation is, in fact, an important attribute that reflects NPO's flexibility and durability as a practical data-gathering tool.
But the tendency to participate has to be handled carefully by the non-participative observer if it is to remain an attribute. The same can be said for the related tendency of becoming more noticeable in the setting. This can undermine the ‘naturalness’ of the behaviour under study, as those being observed are reminded of the researcher's presence. It is difficult, if not impossible, for the non-participative observer to maintain the ideal of invisibility, of being a ‘fly on the wall’. I did not, and could not, attempt to become a fly on the wall when observing. While observing I sat in a small room with the dozen or so members of the residents' committee. In such situations, it is more helpful to abandon the ideal of invisibility. Instead, the observer can aim to become a familiar presence in the setting, a presence that excites no comment and that does not affect behaviour. An alternative, one that is at the same time possible to sustain and is compatible with the non-participative role, is that of being akin to a piece of furniture: visible but familiar. Over a period of months, I had taken time to make myself known to the committee and to establish and develop rapport so that I was not a complete stranger when I started to observe the meetings. I was seldom, if ever, an unseen fly on the wall. Instead, my presence in the meeting room, from the perspective of the committee members, can be characterised as familiar and routine.
In practice, tendencies towards participation and visibility do not undermine the effectiveness of NPO as a data-gathering tool. As in all research methods, practical application can deviate from ideals. Irrespective of the method that is used, acknowledging the challenges inherent in data collection requires flexibility on the part of the researcher. NPO is able to meet the demands of practical research.
Non-participative observation: the ‘richer’ relation?
Increasingly, political researchers are using data-gathering tools in combination, either sequentially or consecutively (Blaikie, 2000, p. 270). Here there is increased potential for the employment of NPO, and its particular attributes, as part of a multi-method data-gathering strategy. In my research, NPO is utilised alongside interviews, questionnaires and documentary analysis for the checking and comparison of data, and for highlighting interesting inconsistencies. Crewe adopts NPO (as well as interviews and questionnaires) to supplement data collected through PO. However, Crewe does not explicitly acknowledge her adoption of NPO, even though she discusses her use of interviews and questionnaires in relation to her research project. In spite of such an omission, it is still possible to perceive that Crewe uses it to ‘fill gaps’ left by the other data collection tools (Bryman, 2004, p. 458; Crewe, 2005, pp. 243–247). This demonstrates both the current status of NPO in political research – stressed at the beginning of this article – and its usefulness.
Rather then suggest that NPO is a method that all political researchers ought to adopt, I instead urge a wider acknowledgement and consideration of NPO when planning data collection strategies. This will entail political researchers paying more attention to methods and methodological issues, making space for more sophisticated and challenging approaches to data collection.
NPO has a particular set of attributes that merit it being a member of the family of research methods routinely considered and employed in empirical political research. However, its position and relationship to other data-gathering methods will remain unclear until political researchers – and authors of political research texts – recognise and appreciate more fully those attributes and, where appropriate, incorporate it in their data-gathering strategies.
Footnotes
I wish to thank two anonymous reviewers for their comments on an earlier version of this article.
1
For political researchers interested in non-participative observation, research methods literature from other academic disciplines offers comprehensive introductions to the approach. Education and psychology research methods texts, in particular, are worth exploring.
2
Mossbank refers not to the town but to an area (or neighbourhood) of the town.
3
These include covert and overt, structured and unstructured and simple and contrived approaches to observation (Bryman, 2004, p. 167).
4
Prompts and notes from a non-participative observation session are reproduced in Wilkinson and Birmingham (2003, pp. 135–141).