How to analyse qualitative data

Qualitative data comes in a number of different forms, so there is no ‘one-size fits all’ answer. This article provides an overview of the different research methods and the data gathered, with options for how to practically go about your analysis of qualitative data in research.

What is qualitative data?

Qualitative data collection is in-depth and focussed on a smaller sample size than quantitative data. Undertaken usually in-person with individuals or in small groups, it investigates perceptions, beliefs, attitudes, reasons why decisions are made, and aims to understand and observe behaviour. Qualitative research output is based on interpretation and is expressed in language, images, or even sound bites and video.

Methods of qualitative data collection

Qualitative data can be collected in a broad number of ways, both via direct questioning, observation, and recording by the subject themselves. Data can be collected proactively from individuals or groups, or by gathering data shared via social media. If interviewing a set of respondents (either one-to-one interviews or in groups), a sampling method will be applied to get a cross-section of the target audience (e.g. purchasers, non-purchasers, age, gender, region).

Qualitative data can be collected via;

• In-depth interview (IDI)
• Focus group
• Online forums and communities
• In-home product testing
• Diary
• Ethnographic observation/‘Shopalongs’
• Social listening
• Survey/questionnaire (Open-ended questions within quant survey)

Qualitative data formats

From such a range of data collection methods, it’s no surprise that you can produce huge volumes of data, in a range of different formats.Data can consist of text, both verbally delivered or written, or can be a set of visual images recording activity, behaviour or emotional observations in real, observed situations e.g. reaction to new product, actions when shopping. Individual interviews and groups are usually recorded with a transcript provided.

Qualitative data analysis methods

Different interpretation approaches are used to analyse qualitative data depending on the objective, whether it’s understanding specific content elements such as mention or recall of brands in content analysis, the stories people tell using narrative analysis, themes and topics, such as from those providing feedback on a service within thematic analysis, the development of grounded theory or understanding customer norms or ideas within discourse analysis.

Preparing the data for analysis

The key challenge of qualitative data analysis, whatever the objective, is the broad range of data collection techniques and multiple formats of results: how is it possible to bring all this together to create a cohesive set of conclusions?

Qualitative data analysis needs to enable;

1. organisation of multiple methods of research (e.g. IDIs, desk research, text documents) across multiple data sources (individual respondents, groups of respondents, websites, social media platforms) and formats (e.g. images, text, behaviours), including text heavy datasets
2. coding and the creation of research themes

1.Organising the data

The sheer volume of data can be difficult to sift through to uncover themes.Various tools are available to help summarise and filter the data;

• Spreadsheets can be used to begin to organise the data from different sources, as well as record content of images and video
• Mindmaps enable you analyse themes in a visual way
• Text analysis software can display wordclouds and uncover themes
• Qualitative analysis software that can manage the different inputs effectively
• Coding software can organise text data into a structured format for analysis

2.Coding

The text heavy nature of qualitative data means that by reading large volumes alone, it’s impossible to get a balanced sense of the overall themes without some form of quantification. This quantification is done with a process called coding. Coding enables you to group open-ended data into broad or narrow themes which can be applied to the whole dataset in a structured way. The data can then be quantified into these themes or categories and even analysed by the different respondent profiles. We cover separately the seven stages of analysing open-ended question data using the concept of coding.

Coding can be done manually either in Excel, within a qualitative data analysis package or using a dedicated platform such as codeit. Although Excel is available widely, it is a poor choice for this kind of large scale qualitative data, so we compare the other platforms below.

Qualitative data analysis software

There is no one-size-fits all for qualitative analysis, with the options falling broadly into three categories;

1. Single source text answers on a range of pre-determined topics, such as survey open-ended responses
2. Multi-source text, across mixed topics or sources
3. Other non-text based data sources, across various formats such as image, behaviour or video

For single-source text answers, it’s possible to use a dedicated coding platform, such as codeit. For multi-source text and non-text data, there are dedicated qualitative analysis platforms designed to assist.

Open-ended coding software

Dedicated coding software tools such as codeit are designed specifically for open-ended survey questions, with multiple functions built in to make the process efficient, fast and cost effective. Such tools enable accurate coding to be completed fast, including the generation of topic themes within seconds. Further, for ongoing studies of large volumes, Machine Learning is harnessed to automate further coding based on the initial human-curated data.

Qualitative-specific analysis platforms

For studies encompassing images, behaviour observation and research across multiple sources, dedicated platforms exist to assist with the organisation of analysis and coding across differently structured data sets and sources.

The table below compares the two options;

If your survey is single source, with open-ended responses, you can find out more about how to code by reading the seven stages of analysing open-ended questions and the practical steps of how to analyse open ended questions.

Key takeaways

If you have multiple qualitative data sources and formats, you’ll need to use a dedicated qualitative software for your analysis. However, if the majority of your data is text-based answers from a small number of different sources, you can consider using a dedicated coding platform such as codeit which will give you all the power of coding, with results available fast in a structured format of themes, enabling easy, actionable, research findings.

If you’d like to give codeit a go, you can take up their 30 day free trial to help you start coding and analysing your open-ended question data today.