Here are notes from the opening lectures, very much scribbled down on the iPad. Lots of interesting discussions around our potential research projects, and I left feeling like I wanted to explore critical discourse analysis in more depth. I think approaching the model music curriculum through this lens will be useful.
Quantitative methods:
Numerical data
Large observations
Basic raw data
Surveys, structured obs and content analysis
Explanation through rigorous measurement
Allied to positivism, objectivism and falsification
Ontologically social = disconnect
Quantitative methods utilise a deductive approach (Bryman, 2008).
Order of approach;
Theory
Hypothesis
Design
Measure of concepts
Space and subjects
Collect data
Process data
Analyse data
Findings/conclusions
Write up
Build on theory
Key terms of quantitative research = operationalisation, precision, consistency. Helps organise ideas and concepts.
Surveys use left-right attitudinal scale; five point scale
Encouraged to use pre-existing data.
Causal inferences including dependant and independent variables.
Validity = does it measure what we say. Relationships between variables and all are included.
Reliability = is it consistent, intertwined with validity.
Generalisability.
Replication , should be able to be carried out by any other researcher. Objective
Qualitative research
textual or visual
Smaller cases, highly complex/messy data
Interpretivist
Interviews focus groups, document analysis
Rejection of quantitative research terminology
Scientistic, exaggerating value of those approaches
Unavoidably subjective
Embrace the bias
King, Keohane and Vane book
Approach = inductive stages;
Research questions
Selection of subjects
Collection of data
Interpretation
Conceptual developments
Writing up of findings
Quant v Qual sectarianism is a very real thing, with roots deeply embedded in epistemology and ontological.
Pragmatism wins due to RQs, available data, level of analysis, audience
Mixed methods
Utilising strengths of both to service the research questions
Triangulation
Completeness, gaps filled
Exploring difficult stuff so should use the tolls available
Credibility = quants more respected in certain arenas
Methods should be methodically thought out with regards to RQs.
Asking questions in quantitative and qualitative research
Surveys
Quantitative research, positivist, deductive
Experimental designs
Survey mode = delivery of survey
Questionnaire interviews/structured - disadvantages = resources, social desirability, differential interviewer effects
Paper self-completion questionnaires: no social desirability, no interviewer variability, better responses to sensitive questions, resource light. Disadvantages = high level of literacy needed, need to be shorter, lower completion rates.
Online self-completion questionnaire: more attractive formats, routing programmed as automatic, relatively easy, fast and cheap. Disadvantages = practicalities, IT literacy
Asking the right questions: one floor can make entire datasets invalidate; careful design is required.
Harmonisation = has what attempting to measure been measured before? Principles of replicability
Closed questions with clear language.
Define ambiguous terms and phrases.
Qualitative interviews
Jacques-Alain Miller - against surveys
Ann Oakley - surveys are dehumanising, exploitative and hierarchical.
Qual research much more flexible and messier
Encouraged to stray away from script, probe to help understand further. Enables perspectives of participants to emerge, inductive production of theory. Narrative driven.
Unstructured interviews
Semi-structured - most common. Used themes but flexible approach
interview guides
Start general, then zoom in. Naturalistic
Icebreaker usually a good idea
Location is important.
Pilot interview guide
Fast mental reasoning needed (mason, 2002)
Some more methods are complemented as semi structured interviews are often viewed as overused.
Observational methods: structured observations
Anthropologists say Qual interviews are not reliable observations.
Structured observations
A method for systematically observing behaviour of individuals. Leads to quant data. Grounded in positivist theory.
Involves explicitly formulated rules for observation.
Codes represent indicators, need to be clearly focused and closed.
Sampling:
Representative of those being studied, must be random, probability sampling, little bias as possibly
Size doesn’t matter,it’s how we get it. Probability sampling is realistically difficult.
Pros:
Analyses potential gaps between stated and actual behaviour, unconscious habituation.
Scientific approach
Cons:
Consistency
Reactive effect
Coder fatigue
Imposition of the coding framework
Participant observation and ethnography
Generate theory by use of data analysis using Qual methods.
Ethnographic research = immersion in group being studied. Listening and engaging. Developing understanding. Broad social concepts.
Ethnographic field notes extensively. Writing up is part of the process of understanding. Enables better understanding of culture and society.
Participant observation is more modest approach.
Sampling with ethnography and participant obs:
Strategically in relation to RQs = purposive sampling
Convenience sampling = what you can get
Accessing groups can be hard, permission from senior positions. May need to negotiate a research bargain.
Fieldnotes need to be:
Done ASAP, vivid and clear, substantial and numerous continuous. Collection and analysis throughout.
Comentarios