Community Impact of Public Processions

Annex A: Stage 3 Research Methods^[64]

The impact of processions in each live case-study area was examined through a range of qualitative and quantitative research elements both pre, during, and post the procession being investigated.

a) Interviews

Around each of the 'live' Group B processions, semi-structured interviews were conducted with key stakeholders. These interviews explored the organisers' aims and plans for the event and were also used to agree the other elements of this field work stage where possible. These interviews (with event organisers, local authority representatives, senior police officers) sought to identify the preparations that were taking place prior to the procession, purpose and significance of the event, plans for stewarding and collaboration with local communities, anticipated numbers, travel arrangements and any potential concerns and (if appropriate) how these would be addressed. These interviews considered the events but also provided an opportunity to discuss field-work issues such as the facilitation of the ethnographic element of the study, location of structured observers on the day, ensuring organisers, police and stewards were aware of the research team and where they were located. In addition to the Group B sites, interviews with police and local authority stakeholders also covered the SDL and Airdrie processions.

b) Focus groups

Focus groups were held for selected processions aimed at identifying participants' understandings of the aims of the event, its cultural significance, the impact of the event on the community and if relevant, formal and informal responses to this.

Focus groups varied in size (from seven - 24 respondents)

• Two Orange Order focus groups - one male, one female
• One Irish Republican focus group
• Community Council - Parkhead
• Community council - Pollokshields
• Police - Coatbridge/Airdrie
• Three key stakeholder groups (local authority and Police Scotland representatives).

Community representatives tended to have a formal role in representing particular interest groups within the local community in an 'official' capacity. Particular attention was given to ensuring representation from different community groups (in terms of age, gender and ethnicity/religion). A member of the research team attended focus groups organised by the Scottish Women's Convention and Glasgow Women's Library (Mixing the Colours) which had been set up to discuss women's experiences of sectarianism and where issues of public processions were raised.

c) Ethnographic and structured observations

An important part of this study involved participating in and observing processions as they happened. On the day of the procession, the actual event was carefully researched. The opportunity to walk alongside processions and to observe the impact of the procession as it passed along the route provided a number of important contexts for data collection and analysis. As more recent crowd psychology research has emphasised, crowds of people can and do have a social identity, and can police that social identity quite strongly (Stott et al, 2008a and 2008b). At the same time, groups represent a mixture of social and individual identities, challenging any notion that they are homogenous entities. Ethnographic participation and observation also provided an opportunity to obtain the perspective of the participants and what the event 'means' to them. Similarly, in order to understand the 'impact' of the procession requires an opportunity to obtain the views of those immediately affected.

During the events selected for in-depth analysis, structured observation was carried out with paired researchers who observed the procession as it occurred, from separate points along the route. Using a checklist approach (as used by Stott et al., 2008 in football crowd research) the researchers charted the progress of the procession, noting a range of characteristics and identifying any changes over time^[65]. Researchers took notes around predefined categories at agreed points in time (generally every fifteen minutes), focusing upon, for example the nature of police deployment and the quality and quantity of interactions between participants, spectators and police. This method of 'scan sampling' has proved effective for measuring behavioural states and incidents which occur, and police responses. The location of the researchers was identified in discussion with police and procession organisers, and in the case of the July 6^th Coatbridge procession during attendance at a pre-event briefing meeting. At this meeting, detailed plans of the route of the procession were discussed and the priorities for police and stewards (in terms of public safety and anti-social behaviour) were planned and agreed. The position of researchers when observing processions is noted by Komarova and McKnight (2013).

In addition to structured observations, two researchers undertook walking ethnography and guided observation during the live processions. Unstructured observations provided an opportunity to examine action and reaction surrounding the following:

• (visual) Procession participants dress, presentation, flags, banners - what is the symbolism for the participants and how is this perceived by the observers?
  e.g. links to local communities, industries, traditions; uniforms or insignia. Gender and age of participants and observers; children present - participating/observing? Who are the observers?

• (audible) Music and tunes, singing, shouting e.g. beat of drums - does it alter at different points, volume - increase or decrease at particular points on the route? Cessation of music i.e. when passing churches, cenotaphs? Singing? If so, what and how? Shouting - from participants or observers? Intimidating or supportive/participative?

• (atmosphere) Ambiance of participants and observers, interactions with crowd, police and stewards, experiential encounters of researcher
  e.g. expressions of smiling, concern; any apparent tensions; family atmosphere?

Where feasible, one researcher was embedded in the procession, walking alongside a representative (or steward) from the organising group. This approach ensured that 'independent' observation, unattached to the event or participants, was supplemented with 'informed' and guided observation by an event 'insider'. The method of 'walking with', and sharing the same visual field as participants has a precedent (Lee and Ingold, 2006; Pink, 2007). The unattached researcher walked along the periphery of the route undertaking a visual tour of the procession, walking with supporters and encountering public bystanders as the procession moved through the town/city. This was supplemented by researcher-created documentary photographs which captured the event visually. Visual tours combining photography by participants, with a tour between participant and researcher where a narrative is produced 'addresses the encounter with the city as phenomenological as well as discursive' (Peyrefitte 2012: 3). Photography was focused on capturing symbolism and imagery associated with the event (i.e. flags, banners, costume, density of crowd) rather than individuals. Specific consideration of the ethical issues related to visual research has been set out (Wilkes et al, 2008) and was considered fully throughout this element of the study.

Research tools were designed to guide ethnographic observations in each site and the same fieldworkers were used across research sites wherever possible. Field notes constituted the primary data source and included the observers' qualitative impressions of the behaviour of participants and spectators, police deployment, interactions between participants, spectators and police, songs and chants, as well as descriptions of specific events, places and people. The element of team work which underpinned the ethnographic research ensured that a number of different dimensions of the events were captured; however it also highlighted the significance which individuals bring to understanding and interpreting situations which they are a part of (Malacrida, 2007). Researchers also remained in the vicinity of the procession after it had ended, providing an opportunity to assess the dispersal (of processors, spectators and police) and any subsequent aftermath.

In seeking to examine the community impact of processions and parades, the study sought to balance the impact of these events on the local communities through which they passed, while remaining cognisant that procession participants were also part of the local community.

d) Scorecard mini-survey and business surveys

As well as conducting observational research around the procession, the experience of the general public around the perimeter of the procession was captured through short scorecard mini-surveys, administered during and immediately after the processional event (covering a three hour time slot in total). These short street surveys took five-eight minutes to complete and researchers targeted spectators leaving the procession, passers-by and people using areas that were near to the procession, or significant in terms of being a key hub for people going to and from the procession (i.e. railway stations and retail premises near the procession).

In addition the survey was amended to allow for the interviewing of business representatives near the survey route at the same time. This allowed us to capture the perspectives of businesses, in particular retailers, on the immediate impact of the procession, and in particular the impact on footfall and turnover.

e) Residential survey in communities bordering procession routes

A residential survey was carried out to directly assess how community members experience processions passing through their community. Processions typically follow relatively linear routes, passing through a number of communities prior to a final congregation point. In surveying whether a procession has had an impact on a community one is typically caught between two options. On the one hand, it is possible to opportunistically sample closely along the route of a procession, maximising the probability that community members who are affected by the procession will be selected, but thereby missing out on drawing a more systematic sample that would be generalisable to the community at large. Conversely sampling on the basis of a broader geographic area or telephone sampling around area codes, typically results in areas that are too large to allow clear discernment of procession impact.

Sample areas were identified via the Postcode Address File (PAF) with these areas being selected using radiuses derived from several points along the procession routes. The addresses were then randomly selected from each of these areas. All selected households were sent a letter in advance of the survey, describing the research and informing residents we would be conducting interviews on a certain date in their area.

Qualitative and quantitative data analysis techniques were used to analyse the data and the research team worked together to ensure systematic and consistent quality analysis standards across the different methodological approaches employed.

In the field, our residential survey approach proved problematic. In Coatbridge we achieved reasonable responses from residents who were at home, but the time of year (the start of the summer holidays) and the fine weather, did result in a high number of householders being un-obtainable. In our subsequent fieldwork areas (Govan, Bridgeton and Parkhead) the logistical difficulties of conducing door to door surveys proved insurmountable given the very high prevalence of access controlled flats and tenements, often with disabled buzzer systems, and with a large number of derelict or vacant properties. Our fieldwork resources and the time available before each procession (usually about ten days), necessitated an alternative approach. We therefore adapted the survey to be administered as a street survey, and deployed fieldworkers at different times and locations before the processions to collect survey material. People stopped in the street who were not resident in the immediate geographic areas were thanked, but were excluded from completing the survey.

Though telephone details, and telephone follow-up surveys were subsequently conducted, gaining people's agreement to disclose their telephone number was much more difficult in an on-street setting. We therefore added an additional survey element, sending out postal surveys to our original, randomly selected households taken from the PAF. These surveys were sent out after the procession^[66].

Though street and postal surveys cannot be viewed as non-random, or statistically representative of the communities in which the surveys were conducted, Table AA1 below nevertheless provides some indication of representativeness.

Table AA1: Characteristics of pre and post survey samples relative to ward areas

Pre-procession Surveys*	Govan (Govan ward**)	Coatbridge (North and Glenboig ward)	Parkhead (Shettlestone ward)
Census % Male	50%	49%	50%
Survey % Male (& no.)***	57% (n=25)	33% (n=15)	52% (n=43)
Census % Female	50%	51%	50%
Survey % Female (& no.)	43% (n=19)	67% (n=30)	48% (n=40)
% Population aged 16 to 29	25%	18%	19.9%
Survey % (&no.) aged 16 to 29	19.3% (n=8)	6% (n=2)	18.2% (n=16)
% Population aged 30 to 44	24%	20%	18%
Survey % (&no.) aged 30 to 44	29.5% (n=13)	32.4% (n=11)	10.8% (n=9)
% Population aged 45 to 64	22%	26%	29%
Survey % (&no.) aged 45 to 64	36.4% (n=16)	44.1% (n=15)	50.6% (n=42)
% Population aged 65+	14%	17%	18%
Survey % (&no.) aged 65+	15.9% (n=7)	17.6% (n=6)	19.3% (n=16)
Post-procession Surveys****	Govan (Govan ward)	Coatbridge (North and Glenboig ward)	Parkhead (Shettlestone ward)
Census % Male	50%	n/a	50%
Survey % Male (& no.)	45.8% (n=38)	n/a	47.4% (n=37)
Census % Female	50%	n/a	50%
Survey % Female (& no.)	54.2% (n=45)	n/a	52.6% (n=41)
% Population aged 16 to 29	25%	n/a	19.9%
Survey % (&no.) aged 16 to 29	12.6% (n=11)	n/a	16.5% (n=13)
% Population aged 30 to 44	24%	n/a	18%
Survey % (&no.) aged 30 to 44	28.7% (n=25)	n/a	34.2% (n=27)
% Population aged 45 to 64	22%	n/a	29%
Survey % (&no.) aged 45 to 64	36.8% (n=32)	n/a	27.8% (n=22)
% Population aged 65+	14%	n/a	18%
Survey % (&no.) aged 65+	8% (n=7)	n/a	15.2% (n=12)

*Pre-procession surveys consisted of street-based surveys in Govan and Parkhead and door to door residential surveys (based on a systematic sample drawn from the Post Office Address File) in Coatbridge.

** Ward figures are based on Scottish Neighbourhood Statistics figures, or on council summaries of those figures. Ward areas in all cases cover a significant portion of the areas surveyed, though in no instance are they simply coterminous with the survey areas. They are nevertheless the best available fit for which relevant Census statistics are available.

***Gender missing for 2 respondents in Govan and for 3 respondents in Parkhead. All similar figures and percentages presented in Table AA1 exclude figures where either there was no response or where respondents refused to provide a response, though the number of these cases was very low.

**** Post-procession surveys were administered via post in Govan and Parkhead. In Coatbridge the majority of post-procession surveys entailed telephone follow-up surveys of pre-procession respondents.

Survey characteristics and possible sources of bias

Looking at the figures in Table AA1, the distribution of survey respondent characteristics, with a few notable exceptions, do appear to correspond reasonably well to the characteristics of the broader ward populations. There are moreover, some clear reasons for some of the exceptions.

Superficially - in the pre-procession surveys - the street based approach appears to have generated a more balanced sample of respondents than the door to door sample. This however, may reflect the fact that the door to door interviewing was conducted during daylight hours and may therefore have naturally resulted in a higher representation of female respondents. Moreover, the Coatbridge survey area focussed on only one portion of the wider ward, and was notably dominated by relatively affluent semi-detached and Victoria villa-type housing, and this may account for the under-representation of young people relative to the population of younger people for the ward as a whole.

There is a tendency for 'bunching' in terms of the age of respondents, particularly with the post-procession postal survey. Though this bias must be borne in mind, one might reasonably expect some bunching towards the middle age groups for a postal survey of this type (see for instance Sheik, K. and Mattingly, S. (1981) Investigating non-response bias in mail surveys, in Journal of Epidemiology and Community Health, 35, 293-296).

Race, ethnicity and survey respondents

Respondents were allowed to 'self-describe' their ethnicity and this led to a wide range of responses, nevertheless we would estimate that no more than 2% of responses equated to 'non-white' categories as found in the general census. This is a very low figure, though entirely in line with the very low non-white population estimates for two of the three areas (1.3% for Coatbridge and 2% for Shettlestone). However, Govan does have a more sizeable ethnic minority population (some 10% of the Govan ward being estimated as 'non-white' in 2010 (Glasgow City Council Estimate). However, the area in which the street surveys were undertaken - and indeed where postal survey were sent - notably excluded the principal areas of Govan which house non-white residents. This was a direct result of our sampling logic, which focussed on the start, and middle parts of the chosen procession route, as the start and middle areas were those noted for friction.

Combining data sources

Extensive ethnographic data collection (including audio and visual recording of processions, social media coverage of processions - which added an important dimension by allowing us to examine the presentation and debates surrounding a particular event on social network forums, walking ethnographic participation and observation, and structured observations) has provided a rich and innovative dimension to the study. The juxtaposition of ethnographic analysis alongside semi-structured interviews, focus groups, the collection of documentary records (procession plans, objection letters etc.) and statistical data obtained from residential door-to-door, postal and telephone surveys; and on-street surveys, has afforded us a comprehensive overview of perceptions and experiences of the community impact of processions in Scotland.