Planning your Dissertation Part 6: 'Data Analysis' by Roy Horn

Wednesday, 6 July, 2011

Data Analysis - it's all about finding a good story to tell!

Once you have all the data for your dissertation you can ‘breath a sigh of relief', you should be able to create a dissertation that passes from this point!

Before you start to do any detailed analysis, it is sensible to ‘validate’ the data you have. Errors can creep into the recording process at a number of points, and carrying out some simple checks can ensure that a participant's responses can be entered into the research rather than having to be discarded because of errors. When you are using online surveys, think carefully about making all the questions compulsory. This ensures that participants cannot miss important questions. In ‘paper’ surveys, as each one is handed in visually check that all the important questions have been completed. The other general task, at the point of data entry, is to make a visual check of the answers to ensure that more general errors have not been made.

There is a myriad of possible methods for analyzing data, and I cannot do them justice in this short space. So if you need ideas or information on how to conduct a quantitative or qualitative analysis, consult my textbook on the subject: (link currently not available anymore) Researching and Writing Dissertations (2009), Roy Horn, CIPD: London, pages 141:219.

The two standard packages for quantitative analysis are SPSS and Excel both of these work well and will provide you with a huge output of the raw analysis. The main problem and issue that must be addressed are to find a ‘Data Story’ to tell. The section in your dissertation that analyses the data and represents the analysis is vital to its success. The effort you have expended so far can be lost or dissipated with a weak or illogical data analysis section.

Analysis is often categorized in parts as follows:

• the study of the constituent parts and the interrelationship of the parts
• the breaking down and separation of the whole into constituent parts
• simplifying the whole into parts to display the logical structure
• an explanation of a process and the parts of that process.

So in creating a Data Story you will be trying to address all these elements. When your dissertation is marked, the marker will be expecting a certain type of story, such as:

1. This is a problem.
2. The problem has these parts.
3. People say these things about the problem.
4. If people did this or that, the problem would lessen or even be solved.

Now take a second look at that story. Does it look like your dissertation? Your data story should:

• be simple, brief and relevant
• have a beginning, middle and end
• have a ‘punchline’, a discovery
• have data and evidence to support the story (but only to support the story)
• be specific and appropriate to the research study
• offer improvement and solutions
• illustrate how it might be relevant to a wider audience.

Be simple, brief, and relevant

Now this immediately presents a problem. How can all your research and effort be simple and brief? The answer is that it must be expressed in this way if it is to be understandable and clear to the reader.  The simple and brief characteristic is achieved by separating out the important outcomes of the analysis. One main idea creates one main story, then the next idea and story, and so on.

Have a beginning, a middle, an end - and a punchline, and evidence

Normal story structure follows this format, and your data stories must follow this format if you want to successfully explain your research. What does this look like for a data story? - Beginning - Analysis showed that (for example) there was a significant difference between the book-buying habits of law students in comparison with business students. - Middle - Explore this in more detail using data and evidence such as tables and graphs. Ask questions about why this might be occurring. Relate the issue to any known theory from the literature review. Create a punchline in the middle section - a ‘wow event’, if you like. - End - Find a conclusion to your story, also a summary - and if the data story presents a problem, say what the solution might be. Relate the conclusion to the punchline. - Data and evidence to support the story - The important thing here is to be sure that the data and evidence do support the story. The data can never be the story.

Be specific and appropriate to the research study

At the beginning of your research you set out some aims, objectives, research questions or hypotheses - all your stories should relate to these. It is best if this relationship is made explicit. One of the marking criteria for your dissertation will be how well you have achieved the aims, objectives, research questions or hypotheses. Making the link explicit in the stories will address that issue in an integrated and effective manner.

Offer improvements and solutions

Some stories have a moral conclusion, and your data stories must have a similar thing. In the data stories you will conclude with improvements, solutions or recommendations. This will allow you a ‘handle’ in the data section that you can return to in the conclusion of your work.

Illustrate the significance to a wider audience

Your data stories are specific to your research, but depending on the philosophical stance your research takes, may have useful implications for groups beyond your context. Even specific qualitative data may have general points that are worthy of mention.

Good planning makes for a successful dissertation! Check out this ‘Tom’s Planner’ schedule with all the parts of a standard dissertation added.

Well! That is the end of the blogs on dissertations. I hope you have enjoyed them and that they helped you to complete your dissertation. Remember! Good organization will allow you to enjoy your dissertation and to be successful. Good luck! Get a head start on your dissertation by using this template and start planning now! This schedule will save you lots of time and energy.