Note: this post is aimed as a memory mnemonic for me not you (I’m studying an MSc in Digital Marketing )
Week 7 – Findings and Analysis
Findings is worth 40% (+ ‘conclusions and recommendations’ worth 10% that directly relates to the findings) of the dissertation hence focus on this area.
- Relate your findings to your objectives.
- A popular way of dividing findings is by including different sections that answer specific objectives.
- Use a range of written paragraphs, tables and graphs to answer your objectives. A mixture of these tools will make your work read better.
Analysis relies on Triangulation…
- You need to analyse and discuss your findings in relation to the past literature. Does it agree/disagree with past findings. i.e. known as Triangulation.
- Triangulation gives research more depth & analysis, rather than general boring SPSS/Excel tables with no discussion.
- They are not just after tables of statistics with little interpretation.
Data & Graphs
- Some data is better represented by graphs or tables. Look at the story you are trying to tell.
- Quantitative analysis techniques range from creating tables or graphs that show frequency of occurrence through to establishing statistical relationships between variables.
- For quantitative research you are expected to at least you bi-variate cross tabulations.
- These would not be appropriate for qualitative research
If your analyses are to be straightforward and of any value you need to:
- Understand what outcomes you want from your research.
- In particular any specific issues you want to understand or relationships you want to test.
- Think about the analysis tools you will use
- Be familiar with the latest software packages available:
- SPSS for Quantitative research (but do not have to use it and majority do not use it). Survey Monkey does have some basic similar functionality.
- NVIVO for Qualitative Research (again you do not have to use this)
….. For both you need real training on both and they are not simple
- Be aware of and know when to use different statistical techniques.
Data Analysis and Data Types can be split into Categorical = cannot measure numerically, and Quantifiable (what I will be using for questionnaires e.g. Likert scale that measures how many people have a certain attitude of opinion by agreeing or disagreeing with a statement made by the person who made the questionnaire)
Coding discussed (around numerical labels that are used if using a statistical package as they help you analyse the various replies) e.g. on SPSS
Describing the central Tendency / Measures e.g. Mode, Median and Mean
Uni-variate (one variable) and Bi-variate (2 variables) Analysis discussed and how you can present the data. Important you go beyond uni-variate analysis.
- Cross Tabulation
- A powerful analytical tool
- Examines responses in the light of others
- Are there any relationships between variables?
Casual is often steered away from due to the complexity of proving casual relationships.
Testing for Significance – Statistical Tests
Deductive research is to test a hypothesis (what I’m doing) hence very important to test for significance in this case.
- Statistical significance means that a particular difference is unlikely to have occurred by chance or through sampling error.
- Significance testing can only be done when using probability sampling
Then Analysis of Qualitative Results was covered – all depends on the research that is collected and very different to Quantitative analysis. It’s a lot about your interpretative approach. Malhotra came up with a nice loose framework if working in Qualitative research.