Data analysis

• Data from Observations Should be Analysed Carefully
• Quantitative data


  Data obtained by categorising and rating behaviour
• Qualitative data


  Data that may consist of a video or audio recording, or written sets on what the observer has seen
• There must be adequate data sampling to ensure a representative sample of participants' behaviour has been observed
• Language must be used accurately - the words used to describe behaviour should be accurate and appropriate and must have valid operationalised definitions
• Researcher bias must be avoided - e.g. it's not okay to make notes only on events that support the researcher's theories, or to have a biased interpretation of what is observed
• When closed questions are used as part of an interview's structure, quantitative data can be produced by counting the number of participants who replied 'Yes' to a particular question
• When open questions are used, more detailed, qualitative data is obtained
• Context


  The situation in which a participant says something, and the way they are behaving at the time, may be important. It may help the researcher understand why something is said, and give clues about the honesty of a statement.
• The researcher should clearly distinguish what is said by the participant from how they interpret it
• The researcher must avoid bias in selecting what to include from qualitative data (e.g. only including statements that support their ideas)
• The interviewer should be aware of how their feelings about the interviewee could lead to biased interpretations of what they say or how it is later reported
• Like observations and interviews, questionnaires can give you both quantitative and qualitative data
• It's especially important to distinguish the interpretations of the researcher from the statements of the participant, and to be unbiased in selecting what to include in any report on the research
• The analysis of written answers may be especially difficult because the participant is not present to clarify any ambiguities, plus you don't know the context for their answers (e.g. what mood they were in, and so on)
• Thematic analysis


  A form of qualitative analysis that involves making summaries of data and identifying key themes and categories
• Different researchers may read different things into the themes in thematic analysis, so it can be subjective
• Strengths of qualitative analysis


  • Preserves the detail in the data
  • Allows for hypothesis formation during the analysis, leading to new insights
  • Some objectivity can be established by using triangulation - other sources of data are used to check conclusions
• Content analysis


  A way to quantify qualitative data by analysing the data into categories or typologies (e.g. sarcastic remarks, statements about feelings, etc.), quotations, summaries, and so on
• Strengths of content analysis

  • Hypotheses may be developed during this analysis so that they are 'grounded in the data'
  • A representative sample of qualitative data is first collected
  • Coding units are identified to analyse the data
  • Statistical analysis can then be carried out
• Limitations of content analysis

  • Often an individual's judgement is used to define coding units, so they can be subjective
  • Reducing the data to particular coding units removes detail, and the true meaning of things may be lost when taken out of context
• Primary data


  Information collected during a researcher's direct observations of participants, e.g. test results, answers to questionnaires, observation notes
• Secondary data


  Information collected from other studies, which can be used to check the validity of studies or provide evidence to support or discredit a new theory
• Meta-analysis


  Where you analyse the results from loads of different studies and come up with some general conclusions
• Meta-analyses can reduce the problem of sample size, but there are often loads of conflicting results out there, which makes doing a meta-analysis a bit tricky