Download a glossary of terms used in this course.
Data analysis methods are the tools and techniques used to evaluate, explore, interpret, transform and/or model primary or secondary data. They can be used to analyse
Your choice of data analysis methods depends on a number of factors, including your epistemology, ontology, theoretical perspective, research topic, research question and methodology. It also depends on the data collection methods you choose (see the module Data collection methods).
Select each researcher to find out more about qualitative, quantitative and mixed data analysis methods.
Continue on to find out more about qualitative, quantitative and mixed data analysis methods.
- Qualitative
- Mixed
- Quantitative
“I'm studying coping strategies of working street children in Mexico City. I'll give cameras to children to capture, discuss and share their stories. Then I'll interview them, using the photos as points for discussion. To some extent, data collection and analysis will take place simultaneously. ATLAS.ti software will enable me to group, filter, segment, code and hyperlink photos and quotations. I will be able to visualise connections and attach memos, and integrate visual and textual analyses.”
“My study looks at how Muslims are discussed in social media using a corpus-assisted discourse studies approach. Topic modelling will be combined with critical discourse analysis to consider the way the word 'Muslim' is used on forums. I'll use a customised web crawler to download and anonymise content and search the content inductively, using topic modelling (a data mining tool to discover hidden semantic structures). I'll integrate the analyses of quantitative and qualitative data.”
“I'm studying the availability of alcohol in urban spaces in the UK. I'll compile a data set of alcohol outlets using the licensing records of local authorities. Then I'll use the Ordnance Survey AddressBase Premium data set to geolocate every household in the area. I'll use theoretical, statistical and practical models to analyse the data, covering aspects such as proximity, topography, access and boundaries. Validity, reliability and reproducibility will be at the forefront of my analyses.”
Research descriptions (Sameer and Nyah) based on Törnberg & Törnberg (2016) and Fry et al. (2017)
Let us look in a little more detail at how Maya analyses her data using artificial intelligence (AI) tools. She works her way through the following steps:
- She enrols on a university workshop about harnessing generative AI for qualitative analysis. She is introduced to CoLoop, ATLAS.ti, MAXQDA and ChatGPT.
- She chooses to use ATLAS.ti. This is a university-licensed commercial AI tool that meets data protection standards and, at her university, offers encryption for storing data. This also provides user access control to ensure that only authorized individuals can access sensitive information.
- She prepares her data by producing transcripts from audio files and converts them to the required format, before uploading.
- She becomes familiar with her data by reading through and highlighting key phrases, sentences and sections.
- She develops codes, first manually by highlighting and using the 'Create Code' feature, and then automatically by using the 'Auto Code' feature.
- She visualises the coded data using features such as word clouds, code frequency tables and network diagrams.
- She identifies patterns, categorises codes into themes and refines the themes.
- She generates a detailed report and exports data ready for presenting findings.
- She discusses her analysis and findings with her supervisor.
- She reviews, revises and cross-checks her results, taking care to consider how bias might have entered the analysis process.
Maya's example illustrates the process of choosing and using AI tools for her research. Within qualitative data analysis there is a suite of choices available, for a variety of analysis tasks. Information about these can be obtained from the University of Surrey in the UK.
AI is appealing because it allows navigation (through literature or data) using natural language exchanges. This is beneficial for many people, notably those new to processes, neurodivergent people and those operating outside of first language. Despite these advantages, there are also weaknesses and problems with AI use. Indeed, the use of AI is heavily contested. More information about using AI for data analysis, along with ethical issues and debates, can be obtained from the module AI in research in the course Principles of Research Methods.
Some researchers struggle to choose appropriate data analysis methods. However, there are a number of factors that you can consider that will enable you to narrow down the possibilities.
Consider the following conversation between Maya, Sameer and Nyah. Then proceed to the question, select an answer and consider the feedback.
Consider the following conversation between Maya, Sameer and Nyah. Then proceed to the question, choose an answer and consider the feedback.
Sameer: How will you choose your data analysis methods?
Maya: I'm collecting qualitative data so this will influence my choice. Maybe some analysis by hand, some using software.
Nyah: I'm the opposite. I'll collect quantitative data so I'll need to look at statistical software.
Sameer: I'll collect both types of data in a mixed approach.
Maya: Will both fit with your methodology?
Nyah: And help you to answer your research question?
Sameer: Good points. I'll have to think about those.
Nyah: So what software will you use, Maya?
Maya: ATLAS.ti is available at my university. I'll be able to work through documents and images, coding and creating categories. This will help me to raise analytical questions.
Sameer: Do you know how to use the software?
Maya: My university runs training courses and there is a free trial on the software website. It'll take time, but it'll be worth it because it'll save on mechanical analysis tasks.
Nyah: I am looking at statistical packages. I've narrowed my options to SPSS, Stata and R. I'm thinking R might be best because it's open source and runs on a variety of platforms.
Maya: Yes, there are also free online tutorials available to help you and you could enrol on a statistics course here at the university.
Sameer: It's daunting, but it's worth taking time to choose the right software and learn how to use it properly!
What is the best way for the researchers to ensure that their data analysis methods are appropriate to their research approach, methodology and research question?
Maya, Sameer and Nyah should read around the subject to increase their knowledge and understanding. They might find it useful to read books on epistemology, methodology and data analysis techniques, as well as the methodology and data analysis sections of Ph.D. theses, for example. Choices can be discussed and refined through conversations with their supervisor or supervisory team.
If they decide to use data analysis software, they need to consider what software is most suitable for their type of project and methodology, and for the data that they intend to collect. They should weigh up different software, using demonstration or free versions. This will enable them to consider features and functions such as handling and storing data, and searching, querying and visualising data. They will also be able to see how approaches and tasks differ, and assess which software is most suited to the way they like to work.
When considering your options for data analysis methods, think very carefully about how they relate to your data collection methods and your methodology. More information about this can be obtained from the Research methodologies module in the Principles of Research Methods course, and from the Data collection methods, Sampling methodsand Data analysis methods modules in the Data Literacy for Research Methods course. Remember that, for some research projects, data collection and data analysis take place simultaneously.
Ensure that the analysis methods you choose will enable you to answer your research question and are appropriate for the type of research you intend to carry out.