Written by Dr. Hannes Nel
Hello, I am Hannes Nel and I introduce the data analysis process and ways in which to analyse data in this article.
You need to know what the different data analysis methods mean if you are to conduct professional academic research. There are a range of approaches to data analysis and they share a common focus. Initially most of them focus on a close reading and description of the collected data. Over time, they seek to explore, discover, and generate connections and patterns underlying the data.
You would probably need to code the data that you collect before you will be able to link it to the problem statement, problem question or hypothesis for your research. Making use of dedicated computer software would be the most efficient way to do this. However, even if you arrange and structure your data by means of more basic computer software, such as Microsoft Excel or, even more previous century, cards on which you write information, you will still be coding the data.
The fundamentals of data analysis
The way you collect, code and analyse data would largely depend on the purpose of your research. Quantitative and qualitative data analysis are different in many ways. However, the fundamentals of data analysis can mostly be applied to both. In the case of quantitative research, the principles of natural science and the tenets of mathematics can often be added to the fundamentals. Therefore, the fundamentals that I discuss here refer mostly to qualitative research and the narrative parts of quantitative research reports. For our purposes a research report can be a thesis or dissertation.
You should “instinctively” recognise possible codes and groupings by just focusing on the research problem statement or hypothesis. Even so, the following hints, or fundamentals on collecting and analysing data remain more or less the same, regardless of which data analysis method and dedicated computer software you may use:
- Always start by engaging in close, detailed reading of a sample of your data. Close, detailed reading means looking for key, essential, striking, odd, interesting, repetitive things people or texts say or do. Try to identify a pattern, make notes, jot down remarks, etc.
- Always read and systematically code your collection of data. Code key, essential, striking, odd, linked or related and interesting things that are relevant to your research topic. You should use the same code for events, concepts or phenomena that are repeated many times or are similar in terms of one or more characteristics. These codes can be drawn from ideas emerging from your close, detailed reading of your collection of data, as well as from your prior reading of empirical and theoretical works. Review your prior coding practices with each new application of a code and see if what you want to code fits what has gone before. Use the code if it is still relevant or create a new code if the old one is no longer of value for your purposes. You may want to modify your understanding of a code if it can still be of value, even if the original reason why you adopted it changed or has diminished in significance.
- Always reflect on why you have done what you have done. Prepare a document that lists your codes. It might be useful to give some key examples, explain what you are trying to get at, what sort of things should go together under specific codes. Dedicated computer software offers you a multitude of additional functions with which you can sort, arrange, and manipulate objects, concepts, events or phenomena, for example memoranda, quotations, super codes, families, images, etc.
Memoranda can be separate “objects” in their own right that can be linked to any other object.
Quotations are passages of text which have been selected to become free quotations.
Super codes can be queries that typically consists of several combined codes.
And families are clusters of primary documents (PDs)), images that belong together, etc.
- Always review and refine your codes and coding practices. For each code, accumulate all the data to which you gave the code. Ask yourself whether the data and ideas collected under this code are coherent. Also ask yourself what the key properties and dimensions of all the data collected under the code are. Try to combine your initial codes, look for links between them, look for repetitions, exceptions and try to reduce them to key ones. This will often mean shifting from verbatim, descriptive codes to more conceptual, abstract and analytical codes. Keep evaluating, adjusting, altering and modifying your codes and coding practices. Go back over what you have already done and recode it with your new arguments or ideas.
- Always focus on what you feel are the key codes and the relationship between them. Key codes should have a direct bearing on the purpose of your research. Make some judgements about what you feel are the central codes and focus on them. Try to look for links, patterns, associations, arrangements, relationships, sequences, etc.
- Always make notes of the thinking behind why you have done what you have done. Make notes on ideas that emerge before or while you are engaged in coding or reading work related to your research project. Make some diagrams, tables, maps, models that enable you to conceptualise, witness, generate and show connections and relationships between codes.
- Always return to the field with the knowledge you have already gained in mind and let this knowledge modify, guide or shape the data you want to collect next. This should enable you to analyse the data that you collected and sorted, to do some deconstruction and create new knowledge. Creating new knowledge requires deep thinking and thorough background knowledge of the topic of your research.
How data analysis should be approached
When undertaking data analysis, you need to be prepared to be led down novel and unexpected paths, to be open to new interpretations and to be fascinated. Potential ideas can emerge from any quarter – from your reading, your knowledge of the field, engagements with your data, conversations with colleagues or people whom you interview. You need to be open-minded enough to change your preconceived ideas and to let the information change your mind. You also need to listen to and value your intuition. Most importantly, you need to develop the ability to come to logical conclusions from the information at your disposal.
Do not try to twist conclusions on the data that you gather to suit your opinion or preferences. Your computer allows you to return to what you previously wrote and to change it. This will often be necessary if you are to develop scientifically founded new knowledge. Your conclusions and ideas might change repeatedly as you collect new information.
Do not be frustrated if, as you progress with your research, you find that the codes on which you decided initially no longer work. Again, you can easily change your codes on computer or cards. You must do this in the interests of conducting scientific research. You will typically allocate primary codes to the issues that you regard as important and sub-codes to less important data or further elaborations on your main arguments. You can change this and change your coding structure if necessary.
The process of coding requires skill, confidence and a measure of diligence. Pre-coding is advisable, but you still need to accept that the codes that you decided upon in advance will probably change as you work through the data that you collect.
At some point you need to start engaging in a more systematic style of coding. You can work on paper when starting with the coding, although there is no reason why you can’t start to work on computer from the word go, seeing that you can change your codes on computer at any time with relative ease. Besides, you can make backups of your coding on computer. This can be valuable if, at some stage, you discover that your initial or earlier codes work better than the new ones after all. You can then return to a previous backup without having to redo all the work that you already did.
You need to understand how the computer software that you are using works and what it can provide you with. Different software has different purposes and ways in which codes can be used. It serves no purpose claiming to have used a particular software if you do not really understand how it works, how you should use it and what it can offer you. Previous students will not always be able to teach you the software because most of the software is rewritten all the time. Rather do a formal course on the latest version of the software that you wish to use.
Most data analysis methods share a common focus.
Data analysis is simplified by coding the data and making use of dedicated computer software.
You can also use coding with simple data analysis methods, for example Microsoft Excel or a card system.
The fundamentals of data analysis apply to qualitative and quantitative research.
You should code data by focusing on the purpose of your research and the research problem statement, question or hypothesis.
The following are the fundamentals of data analysis through coding: Always:
- Start by engaging in close, detailed reading of a sample of your data.
- Read and systematically code your collection of data.
- Reflect on why you have done what you have done.
- Review and refine your codes and coding practices.
- Focus on what you feel are the key codes and the relationship between them.
- Make notes of the thinking behind why you have done what you have done.
- And always return to the field with the knowledge you have already gained in mind and let this knowledge modify, guide or shape the data you want to collect next.
In addition to the fundamentals, you should also adhere to the following requirements for the analysis and coding of data:
- Be flexible and keep an open mind.
- Learn how to come to objective and logical conclusions from the data that you analyse.
- Change your codes at any stage during your research if it becomes necessary.
- Develop your data analysis coding skills, confidence and diligence.
- Acquire a good understanding of the computer software that you will use for data analysis.
- Work systematically.
You will use the fundamentals of data analysis and coding with most data analysis methods.
Almost all recent dedicated data analysis software use coding.
I will discuss the following analysis methods in my next seven or eight videos:
- Analytical induction.
- Biographical analysis.
- Comparative analysis.
- Content analysis.
- Conversation and discourse analysis.
- Elementary analysis.
- Ethnographic analysis.
- Inductive thematic analysis (ITA).
- Narrative analysis.
- Retrospective analysis.
- Schema analysis.
- Situational analysis.
- Textual analysis.
- Thematic analysis.