5.2 Qualitative interviews with women
5.2.4 Data analysis: content analysis and coding
Social research and qualitative data analysis have often been described as “messy”
(Bryman, 2012:15) and filled with “complexity” (Coffey and Atkinson, 1996:4).
Data analysis is not always straightforward, and researchers should explore their data from a variety of perspectives (Coffey and Atkinson, 1996). Qualitative data has been described as being “sexy” with rich descriptions and explanations (Miles and Huberman, 1994:1), and the data analysis process as “art” that requires intuition (Ibid.). In other words, qualitative data analysis requires a careful consideration of the right approach, capable of encompassing all these nuances:
“It is easy to get so wrapped up in one’s data (often gathered at considerable personal cost of time an effort) that one cannot see the forest for the trees and cannot get analytical purchase on the data collected. Too many people are so much in love with their data that they cannot bear to disturb their pristine beauty by interfering with them in any way. Both attitudes are sterile” (Coffey and Atkinson, 1996:2).
The methodology adopted for the data analysis in this study is content analysis.
As explained in Chapter III (Section 3.2.3), content analysis can be defined as a research method that allows the interpretation of data through a process of coding and identifying themes or patterns (Hsieh and Shannon, 2005). Content analysis can be applied to a multitude of data sources including interviews.
Although this study employs qualitative content analysis, providing an in-depth analysis of language and words and how these are being employed by participants, it also looks for the frequency with which each code appears, so that an estimation of the popularity and significance of each code can be established (Ibid.). As in Chapter III, an inductive approach to content analysis was employed in order to avoid preconceived categories. The generation of codes was derived from the data and was guided by RQ3 (i.e. how do women in Greater Manchester conceptualize air pollution and envision a city with clean air?) (Hsieh and Shannon, 2005;
Saldana, 2009).
As explained in Chapter III (Section 3.2.3), coding involves assigning a summative and salient word or short phrase to a fragment of data (Saldana, 2009). Coding in this study was performed in two different ways: (1) as a way of indexing data, so that fragments of data that have a common code can be looked at together and to measure the relative incidence of the codes; but also (2) as a way of analysing data, as patterns and relations were looked at, and categories that grouped codes together were formed. Coding is not just labelling but linking. Therefore, researchers go through multiple cycles of coding. The first cycle tends to be descriptive, an initial approximation of the data, labelling what can be found in the data and condensing meaning into manageable units. The remaining cycles (second, third and so forth) allow linking codes and grouping them in categories or concepts that share some characteristics or meanings (Saldana, 2009).
The interview transcripts were coded using the NVIVO software package (Version:
10.2.2) for qualitative data analysis. For the first coding cycle, a combination of 'descriptive coding' and 'emotion coding' (Saldana, 2009) were used, highlighting the meaning of a passage (such as walking infrastructure) and capturing how participants felt (such as angry or sad). As each new code emerged, the interview transcripts were looked at again to explore if they contained any references to the newly developed codes. The first cycle of coding generated 61 codes. In the second coding cycle (and further coding cycles), attention was paid to the codes and to any patterns or relations between them. Through a 'pattern coding' exercise (Saldana, 2009), codes that belonged together were grouped into categories. In this regard, categories can be considered to be groups of codes that share some characteristics (Ibid.).
For example, the following sentence discusses what measures are needed for clean air, it was first coded as WALKING INFRASTRUCTURE, then under ACTIVE FORMS OF TRANSPORT (this category also grouped with other codes such as SAFE CYCLING INFRASTRUCTURE or CYCLING PROFICIENCY), then under TRANSPORT (this category is also grouped with other codes such as CAR or PUBLIC TRANSPORT), and then under DIRECT FACTOR (because transport has a direct impact on air pollution, in opposition to the other category INDIRECT FACTOR).
I guess it would be pedestrianizing the city centre… having kind of areas that would be exclusively for the people who are walking (Cleanairbex)
First cycle code: WALKING INFRASTRUCTURE Second cycle code: ACTIVE FORMS OF TRANSPORT Third cycle code: TRANSPORT
Fourth cycle code: DIRECT FACTOR
In order to guide the coding process and make it more transparent a codebook was developed (DeCuir-Gunby et al., 2011). As explained in Chapter III (Section 3.2.3), a codebook is a compilation of codes and definitions that is key to analysing qualitative data as it provides a “formalized operationalization of the codes”
(DeCuir-Gunby et al., 2011:138). The codebook generated in this study can be found in Table 5.2, 5.3, 5.4, and 5.5, with each table broken down into different interview questions. Table 5.2 shows the codes generated based on definitions of air pollution and clean air. Table 5.3 shows the codes generated from feelings in relation to air pollution and clean air. Table 5.4 shows the codes generated from motivations for using different forms of transport. And, Table 5.5 shows the codes generated from the measures to be implemented for the transformation to clean air. These tables present the codes, a description of what these codes represent, and the categories which these codes were later grouped into. An in-depth explanation with examples from interviews of the codes and categories is provided in Section 5.3.
Table 5.2 Codes originating from the data transcripts in relation to the two interview questions asking participants to describe air pollution and a city with clean air.
Code Description Category
Sensory 26 References to experiencing air pollution through the senses, to how it can be smelled, tasted, seen, or felt.
Physical
References to air pollution being a grey/black smog or yellow haze that can be seen in and around the city.
Movement 10
References to air pollution being air that cannot flow, that is trapped in the city.
Weight 4 References to air pollution being heavy or oppressive, hanging above citizens, applying pressure.
Dust 2 References to air pollution being present in the form of dust on surfaces.
Dirty air 10 References to air pollution being dirty air filled with chemicals Air
Invisible 6 References to air pollution being invisible, difficult to grasp. Invisible 6 Green / blue
10
References to an environment that has blue and green elements, such as trees, parks, rivers, or canals.
Urban environment
27
Description of a clean air city People 8 References to an environment that is designed for people and not 45
for cars, where people can walk and gather.
Clean surrounding
4
References to an environment that is free of litter, waste or filth.
Scandinavia 3
References to an environment that resembles Scandinavian cities.
Calm 2 References to an environment that is tranquil, where people can walk and be relaxed without the danger or noise of cars.
Fresh air 10 References to air that is breathable, that is healthy or not harmful. Air Alive 8 References to air that has life in it instead of particles, as well as 18
wildlife in the city such as birds or butterflies.
Table 5.3 Codes originating from the data transcripts in relation to the two interview questions asking participants to explain how they feel about air pollution and a city with clean air.
Code Description Category
Worried/concerned 12 References to feeling worried about the health and environmental impacts of air pollution. Angry/Annoyed 10 References to feeling angry about air pollution towards other
citizens, authorities or industry.
Frustrated / Sad 3 References to feeling sad that air pollution exists, that humans caused this problem.
Helpless 3 References to feeling lost about what to do to reduce air pollution.
It could be worse 6 References to other places where air pollution is worse, and to Greater Manchester not being so bad.
It could be worse
6 Pessimistic 11 References to the impossibility of achieving clean air.
Feelings clean air 26 Optimistic 8 References to the possibility of achieving clean air.
Happy 7 References to feeling happy about clean air, to being an ideal situation, a dream come true.
Table 5.4 Codes originated from the data transcripts in relation to the interview question asking participants to explain what forms of transport they use and what their motivations are.
Code Description Category
Efficiency 30 References to arriving on time, travel duration, parking availability, number of connections, frequency of transport.
Motivations for forms of transport
156 Health/fitness 23 References to being fit, being active, being in the right mood, mental health.
Safety 22 References to safety on the road and feelings of insecurity, as well as safety from crime in the streets.
Comfort 20 References to being easy and effortless.
Pleasure 18 References to being enjoyable, pleasurable, that makes one happy.
Weather 16 References to cold, wind, snow or rain, and appropriate cycling clothing and paraphernalia.
Cost 15 References to price, whether it is expensive, cheap or free.
Environment 8 References to being environmentally conscious, not polluting or taking care of the environment.
Disability 4 References to illnesses or physical conditions.
Table 5.5 Codes originated from the data transcripts in relation to the interview question asking participants to explain what measures they would take in order to lead the transformation from a city with air pollution to a city with clean air.
Code Description Category
1
References to fewer cars in the road and parked in the city, to reducing space allocated to cars, and to no car zones.
References to unnecessary car ownership in favour of renting or sharing a car, and to better organized sharing and renting schemes.
Electric / Low emission cars 4
References to electric cars, hybrid cars, or to caps on emissions.
Improved walking infrastructure 12
References to improving walking infrastructure and making streets wider, pedestrianisation, and securing space for walking with prams or wheelchairs.
Active transport
34 Safe cycling
infrastructure 11
References to improving cycling infrastructure and to make it safer, to bike lanes, no tricky junctions, physical boundaries, showers in workplaces.
Chris Boardman 6
References to Chris Boardman and his project for improving cycling and walking infrastructure (Beelines).
Cycling proficiency 5
References to making people more competent in cycling as a form of transport, helping find secure routes.
Network 14 References to improvements in public transport networks including, frequency, capacity, and coverage.
Public transport
34 Price 10 References to making public transport cheaper or
totally free.
Ease 6
References to making public transport easier to use, being able to tap in and out in every transport system, or screens with information about services.
Electric / Low emissions 4
References to making public transport electric in some routes or hybrid in others, having solar panels in roofs of vehicles.
Flexible work hours or place 5
References to being able to work from home or to come and go at different times to avoid daily commutes and rush hour.
Flexible work hours or
place 5
More green and blue 30
References to the need for more parks or green elements in the city such as roofs, wall, trees, grass, flowers, rivers, or canals.
Green and blue areas 54 Management 11
References to maintaining green and blue areas and not letting them overgrow, attract antisocial behaviour and become unusable.
Environmental benefits 9
References to environmental benefits of green spaces such as reducing air pollution and flooding, or cooling effects.
Back gardens 4
References to using back gardens for planting trees and flowers to create a network of greenspace, and to not block pave these for car parking.
More information 6
References to the need for more real-time air pollution information, and more information in relation to actions and available alternatives.
Engagement 16 Responsibility 5 References to citizens claiming ownership and
responsibility, and to bottom-up approaches.
Change in mind-set 3 References to changing the way people think about convenience and material wealth.
Hubs 2 References to the power of workplaces, universities or schools in encouraging change.
Litter 16
References to cleaning the streets from litter and waste, and to creating pleasant environments for
walking. Pleasant spaces
25
Indirect factors
66 Homelessness 9 References to sorting out homelessness and drug use,
and to creating pleasant environments for walking.
Community 17 References to creating strong communities to benefits for citizens and the environment.
Community 17
Health 15
References to how greenspace and active forms of transport improve fitness, well-being, and have therapeutic effects; and how healthy people are more likely to appreciate and care about the environment.
Health 15
Safety 9 References to the need to create a city where everybody feels safe, and to spaces that feel unsafe.
Safety 9
Although coding is widely used as a process for data analysis, there have been methodological objections to it. Two of the most salient criticisms associated with coding are that it is reductionist and that it distances the researcher from the data (Coffey and Atkinson, 1996; Saldana, 2009; Bryman, 2012). This study subscribes to the widely acknowledged counterargument that, looking at it holistically, when codes are employed to form patterns that bring the multiple voices of the participants together in one narrative, coding is not reductionist. Furthermore, in agreement with Saldana (2009) coding does not distance a researcher from the data but brings them closer together. A code, a word that is chosen so carefully
and meticulously, requires spending long periods of time understanding the data.
As Saldana (2009) argues, it is often the case that a researcher remembers not only all of the assigned codes, but also the exact words and phrases used by the participants under that code.