Qualitative Data Analysis Nicola Pugh Data Analyst Birmingham

Qualitative Data Analysis Nicola Pugh Data Analyst Birmingham Public Health

Journey So Far… • Workshops: – – – Focus groups Mapping Questionnaire design Participating risk Active referrals

Definitions • Survey - paper or online questionnaire to collect opinion • Engagement - actively seek public interaction and involvement on areas of need • Consultation - seek public opinion on areas of change

The Gunning Principles … whether or not consultation is a legal requirement, if it is embarked upon, it must be carried out properly. This means that consultation must be undertaken at a time when proposals are still at a formative stage. It must include sufficient reasons for a particular proposal to allow those consulted to give intelligent consideration and an intelligent response. Adequate time must be given to this and the resulting decision must be conscientiously taken into account when the ultimate decision is made.

The basics… Response data falls into two categories Quantitative NUMBERS Quantitative Count of responses to closed questions e. g. yes / no. Usually presented visually as a pie chart or graph. WORDS Qualitative Free-flow text responses from open questions, interviews or workshop discussions. Usually presented as a summary.

What is Qualitative Data Analysis (QDA)? • …a process of collecting qualitative data into some form of explanation, understanding or interpretation of the people and situations we are investigating. • …to examine the meaningful and symbolic content of qualitative data. • …identifying key themes and summarising

Qualitative analysis May come from many different sources • • Short comments/ ‘other’ responses on questionnaires Longer free-text responses on questionnaires Written observations Verbatim text from focus groups and interviews Notes of public meetings Unstructured responses Email, texts, social media feeds, videos and even photographs

Qualitative analysis • Qualitative data is not the same as quantitative data. It is complex and subjective – it is someone’s opinion and therefore cannot be treated in the same way as statistical data. BUT • Just because it is not mathematical data, does not mean that you cannot apply systematic and rigorous analysis techniques. • Qualitative data, more than quantitative, is extremely prone to bias and systematic analysis helps prevent this.

Qualitative Process of Analysis Code text under Themes to be used in the Findings Report CODE data (assign code label to text) Iterative READ through data (get sense of material) PREPARE data for analysis (transcribe field notes, etc) COLLECT data together (online survey results) Cleanse Code the text (descriptive) to be used in Findings Report Simultaneous • • • Collect Prepare Read Code Repeat Organise

Coding techniques • • • Word repetitions – for commonly used words ; these may also indicate emotions Indigenous categories terms used by respondents with a particular meaning and significance in their setting. Key-words-in-context – look for the range of uses of key terms in the phrases and sentences in which they occur. Compare and contrast – Ask, ‘what is this about? ’ and ‘how does it differ from the preceding or following statements? ’ Searching for missing information –try to get an idea of what is not being done or talked out, but which you would have expected to find. Metaphors and analogies – people often use metaphors to when talking about their values and these may indicate the way they feel about things too. Connectors – connections between terms such as causal (‘since’, ‘because’, ‘as’ etc) or logical (‘implies’, ‘means’, ‘is one of’ etc. ) Unmarked text – examine the text that has not been coded as a theme or even not at all. Pawing (i. e. handling) – marking the text and eyeballing or scanning the text. Circle words, underline, use coloured highlighters, run coloured lines down the margins to indicate different meanings and coding. Then look for patterns and significances.

Coding Frame • By organising your key phrases, word or patterns into themes, you have started to create a structure you can apply consistently to all responses. • This is called a Coding Frame. • The Coding Frame should include definitions or explanations about why you have coded something in a particular way. • Consider how you code responses to different questions • Constantly changing and growing - should develop as new codes are added and refined. • Review your Coding Frame regularly and reapply it to responses already coded to ensure consistency.

Flat Coding v Tree Coding Non- Hierarchical coding (flat coding) • A simple list of code • There are no sub-categories. • This works well when analysing survey responses through Be. Heard and ordered by question (i. e. a coding frame for each question) Hierarchical coding (tree coding) • Codes are arranged into parent codes (main code) and sub-category codes (a branch of code that stems from the parent code). • Codes relate to their parents by being 'examples of. . . ', or 'causes of. . . ' and so on. • This type of coding is a good way of showing links between codes and an overarching theme.

Flat Coding v Tree Coding v. Close, generalised friendships v. Sporting friendships v. Sports club members v. Work friends v. Making new friends - same sex v. Making new friends - different sex v. Losing touch with friends v. Becoming sexual relationships

Examples of Coding Key themes: • Own home • Lonely • Independence • Moving out of parents • Conflict • Dependence • Desire for company • Depression • Inactive

Key themes: • Living alone • Relationship with parents • Independence • Relationship with father • Lonely

Quick word about cleaning data Data cleaning is the process of dealing with (or eliminating) invalid data • • Cleaning - You can clean data before data entry (where obvious errors are dealt with) or after data entry (whilst you are coding). Validation - set validation rules to ensure data is cleaned consistently. Example, how do you deal with responses where : • they have misunderstood the question • put answer in the wrong box • double ticked boxes or no boxes at all Rules may include: • Omit the response completely • Re-code the question so that the entry is allowed (e. g. for double ticking) • Include the response elsewhere • • None are perfect, but whatever you choose, apply consistently and highlight in your report Missing data – don’t assume anything from missing data. Field should be left blank. Weighting - Public Health do not apply weighting, but if you do then make this clear in your report.

Summary • Allow plenty of time to code and prepare your report • Analyse responses as you go, don’t leave it to the end • Put comments under headings or themes to build a coding framework • Include comments or reasons for your coding to keep comments in context • Your report should be independent and fairly represent the views of the citizens

Useful sources • University of Huddersfield, Online QDA http: //onlineqda. hud. ac. uk/Intro_QDA/what_is_qda. php • Consultation Institute http: //www. consultationinstitute. org/ • Introduction to Codes & Coding, Chpt 1 http: //www. sagepub. com/upm-data/24614_01_Saldana_Ch_01. pdf • Learn Higher, online resource http: //archive. learnhigher. ac. uk/analysethis/main/qualitative. html • Glaser, B. , and Strauss, A. (1967). The Discovery of Grounded Theory.