CAPÍTULO III DISEÑO METODOLÓGICO
3.4 Hallazgos
3.4.1 Matriz 1 Diarios de campo y observaciones
Possible data choices
It is critical to note the distinction between methodology and method. Methodology is concerned with the study of methods and the way in which knowledge is produced. Each methodological approach will be moulded by specific ontological and epistemological assumptions (Grix, 2004). This is often confused with method,
which is the specific “techniques and procedures used to collect and analyse data”
(Grix, 2004, p.170). In its simplest terms, there are two main types of data used within research, text or written data (qualitative) and numerical data (quantitative). Each has a philosophical rooting and both have their merits and flaws.
Qualitative research is associated with the interpretative approach and methods can include unstructured, semi-structured or structured interviews, participant
observations and focus groups (Gratton and Jones, 2004). This approach “usually requires an extended and intensive period of involvement with the social world”
(Blaikie, 2000, p.242). The focus of this thesis was to gain an understanding of
individual’s perspectives, i.e. their opinions on financial support from SE.
Comparatively, quantitative research is related to the positivist domain and is likely to have limited if any contact with the people being studied (Blaikie, 2000). Methods can take many forms, with experiments and surveys being the most common. Bryman (1988) noted three other methods, official statistics, structured observation and quantitative content analysis.
Quantitative data mainly looks at quantity and quantifiable results, e.g. the changes in club membership figures, pre-and post the London 2012 Games. The main objective of this research was to investigate the impact that the London 2012 Games had on community sport within the UK and specifically NGBs role in leveraging a participation legacy. It was essential to investigate the attitudes of NGB senior staff and to examine the implementation process of the sports participation policy; this allowed the exploration of the participation figures and the mechanisms that highlight
the efficient relationships or challenges faced between the policy creators and strategy formation (SE and NGBs) and delivery agents (VSCs).
Social research- qualitative and quantitative methods
There has been much debate amongst academics, as many believe that these research methods rely on their epistemological assumptions. Many believe quantitative and qualitative research methods sit within two different paradigms, hence it is not possible to combine the two (Bryman, 2012). Bryman (2012) notes that especially in social science there are areas within qualitative and quantitative research that overlap and have commonality. This has led to some suggesting that mixed methods is a third methodological movement, which was originally referred to as
‘triangulation’ to denote the combination of methods (Blaikie, 2010). Gratton and
Jones (2004) argue that qualitative data helps to provide some evidence and support quantitative measurements; for example, quantitative membership figures from NGBs, both at a national and club level can provide useful data about certain elements of their development and progression as a sport. Yet, even though this gives an insight into the
‘what’, it does not allow an in-depth analysis to investigate ‘why’ this change has happened and most importantly ‘how’. Understanding the ‘why’ and ‘how’ through
dialogue with NGB senior managers and club coaches will provide some explanation to the quantitative measurement and can highlight why these participation changes did or did not occur.
Denscombe (2014) states that by combining qualitative and quantitative methods you can recognise and utilise the benefits of both methods, whilst reducing their respective weaknesses. Blending qualitative and quantitative methods can produce a final product that demonstrates the significant contributions of both (Nau, 1995). This was utilised in the current research, with qualitative data primarily collected in study 1, via in-depth interviewees with VSCs. These findings were then used to inform and complement study 2, which consisted of a mainly quantitative, national online survey. This method allowed a large sample of data to be collected and analysed across a wide range of NGBs, which would not have been possible if qualitative methods were used alone. This combination allows more complex research questions to be addressed, than one method could achieve solely and a richer, stronger range of evidence to develop (Yin, 2013).
A mixed methods approach has been selected, as it is the most suitable method to answer the studies’ research questions. Blaikie (2000) states that there are no clear connections between paradigms, research strategies or methods of data collection and analysis. Rather some methods are more commonly associated with a particular research strategy. Hence, issues are limited as long as the researcher is aware of the methods being used and in what context, it is related to particular research strategies (Blaikie, 2000).
Reliability and Validity
Reliability and validity are two key concepts by which the quality of research is assessed, in effect they assess how ‘truthful’ the research is (Gratton and Jones, 2004). Reliability states that the processes within the research, such as data collection methods, can be replicated with the same results (Yin, 2013). There are three main types of reliability; stability, inter-observer reliability and internal reliability. Stability, in simple terms refers to whether the measure is stable across time (test- retest). This ensures that there is confidence in the results, measurement and sample as the results do not fluctuate if tested and then retested (Bryman, 2012). Within this research, the relationship between the research participants and the context in which they were investigated is significant. This context is liable to adapt and vary over time due to staff and funding changes, meaning it was not a requirement of this research for participants to be tested then retested. Therefore, stability did not directly impact the reliability of the results.
Internal reliability relates to whether certain indicators selected by participants tend to relate to their scores on other indicators (Bryman, 2012). For the quantitative
survey the internal reliability was tested using Cronbach’s Alpha. This “essentially
calculates the average of all possible split-half reliability coefficients” (Bryman, 2012, p.170). An Alpha coefficient value will vary between 1, which suggests internal reliability is perfect, to 0 which suggests no internal reliability. Within the literature, the value, 0.80 is typically highlighted to represent an acceptable internal reliability (Bryman, 2012).
Inter-observer reliability is the level to which two researchers would give similar or the same scores to the same social phenomena. As there is a highly subjective element of data translation into categories, a lack of consistency can occur
(Bryman, 2012). For the qualitative elements of the research, which are present in both study 1 and 2, reliability was insured by using inter-coder reliability. This is the degree to which the researcher and two other individuals agree about the coding of items (Bryman, 2012). There are three main threats to reliability (Gratton and Jones, 2004): 1) Subject error. The subject may respond differently depending on the time of day they are asked to supply data, which can lead to bias or error in responses. Within this research this threat to reliability was eradicated as the objective was not to test-retest reliability of participants, as highlighted above. The same sample selection and communication processes were followed for both phases of the research.
2) Researcher error. Two researchers may approach the same data collection in slightly different ways which could affect the results. The creation of the survey and interview script followed a consistent method through which the themes and specific questions were outlined. Additionally, an individual researcher collected and undertook all participant interviews across the four minority sports meaning the process was consistent across all data collected. The survey and interview scripts were both screened prior to distribution by six individuals involved within the community sports sector either practically or as an academic. Then from a participant recruitment perspective individuals were contacted via details available within the public domain or through the human resources department at each NGB. Thus, these steps helped to eliminate reliability error by the researcher.
3) Subject Bias. The participants may provide a response that they believe the
researcher wants to receive that is more socially acceptable or ‘correct’. To
limit the impact of this factor, anonymity and confidentiality was stressed to all participants in both the survey and interviews. Also, the importance of participants providing their own accounts of the topic being investigated was emphasised, reinforcing that there is no right or wrong answer.
Validity is concerned with whether the measure really measures the concept or social phenomena under investigation (Bryman, 2012). Face validity was deemed a suitable measure for validity for both studies 1 and 2, as the script of questions for
both studies was derived from the same triangulated method (investigation of existing literature, exploratory focus groups and document analysis). The survey and interview script was reviewed by a total of six sports sector and academic professionals to test the content, ensuring clarity and understanding. This provided the ability to ensure the two research methods accurately addressed the topic under investigation.
Construct validity relates to whether the constructs used within the research to identify different elements of implementation are an accurate measure of implementation theory. This is relevant for studies 1 and 2, as both include constructs that link to implementation theory. Construct validity will be tested within the data analysis, through comparisons between the existing literature, the data itself and the
researcher’s interpretations, through a triangulation method, which is suggested to
increase construct validity (Yin, 2013). Furthermore, Yin (2013) states that construct validity is obtained when a chain of events is established, i.e. clear links are shown between the theory, research questions, data collection and analysis and the conclusions that are drawn.
Researcher and the research-
It is critical to clarify the role of the researcher, to ensure an understanding of their influence and the level of involvement they did or did not have within the research. With the selected structuralist retroductive research strategy, the role of the researcher is similar to the deductive strategy. The objective is to establish the existence of real structures and mechanisms, using theory-dependent observation (Blaikie, 2000).
The topic area was chosen for a mixture of academic, social and personal reasons. Firstly, from an academic perspective, within the area of legacy, policy implementation and sport there is real need for a wider base of literature to develop to help address the current challenges being faced by the community sports sector. Furthermore, this research provided a greater understanding of how SMEs are being leveraged and the opinions of the stakeholders involved in this process, adding to the limited body of research surrounding empirical leveraging examples. This is supported by Chalip (2014, p.10) who state, “more work is needed…to explore event leverage
in the context of event portfolios”. Personally, sport and the specific topic area of
community sport setting was one of individual curiosity and importance. Then socially, the research was undertaken to contribute to society, by providing findings and greater explanation to aid the improvement of the sports development sector.
It is worth noting that both elements of this research, study 1 and 2 involve NGBs and VSCs as research participants, hence it is important to understand the relationship between these organisations. NGBs are comprised of a range of individuals, from the Chief Executive Officer (CEO) and paid staff at the head office, through to volunteers and coaches who run the local community VSCs. All of these individuals form and determine to a certain degree the action and direction of the organisation. CR accepts the connection between the researcher and the participants within the social phenomena under investigation. Yet, as the researcher is not directly involved with these organisations and is entering externally from a university setting. It was important to consider the impact this could have on the researchers
understanding and analysis, as well as the participant’s responses. Nevertheless,
within critical realism the objective is to explain (by identifying casual mechanisms) how they work and under what circumstances, underpinned by the belief that there is
an external reality “separate to our descriptions of it” (Bryman, 2012, p.29). Hence, the researcher’s objective was to be impartial and conduct a study that would shed
light on the attitudes of specific individuals and the participation mechanisms that were leveraged and implemented within NGBs. This contrasts to interpretivism, where
the “researchers are inextricably part of the social reality being researched” (Grix,
2004, p.83).