STAY UP TO DATE
Join our FREE network for latest news, events, training, research, sponsorship opportunities and more
Data analysis can require specialist evaluation skills. The options for analysing data have been grouped into two areas - quantitative data analysis (numbers) and qualitative data analysis (visual or text).
Quantitative data analysis
For quantitative data analysis, issues of reliability and validity are important:
• You need to show that the methods used were valid and effective in collecting the data needed (e.g. writing non-ambiguous survey questions, using a validated scale).
• You also want to be able to show that the data collection methods are consistent and stable – for example the method would collect the same data regardless of when the data was collected or by whom. A test-retest procedure is used to establish this type of reliability where data is collected at two different times and/or by different people using the same data collection method and the outputs are compared for consistency.
• Eliminating sources of bias introduced by the data collection method, the respondents or the researcher is important. For example, do paper-based surveys yield more responses than online surveys or do online surveys appeal to different sub-groups in the population who may have particular characteristics? If so, can the data be said to be representative of the wider population or is the data biased?
Qualitative data analysis
The first step in qualitative data analysis is presenting the data in a way that can be analysed. For recorded interviews or focus groups this involves creating a written transcript of the interview or focus group.
Thematic analysis is a commonly used process for analysing qualitative data. The process involves identifying and highlighting major themes in written transcripts. This process is called coding. The themes are found within the data but may be influenced by the researcher’s previous reading or experience of the health issue being investigated.
The process of analysing qualitative data is a very personal process. For example, two researchers reviewing an interview transcript could identify different codes and meanings in the content based on their experience, perspectives and knowledge of the subject. A table defining each code is a useful tool to create when analysing qualitative data. For example, in a program investigating attitudes to condom use, the overall theme might be ‘condom use’ with associated codes such as ‘availability’, ‘cost’, ‘experience’, ‘acceptability’ and ‘knowledge’. Participants’ comments about their condom use could then be grouped in relation to these codes. The table defines each code and can be used as a reference tool during the process of coding a number of transcripts about the same topic.
Data analysis software
Data entry and analysis can be time consuming– consider the time taken to enter the responses to a long survey from several hundred respondents! Data analysis software can help you to store, analyse and report your data. There is a wide variety of free and paid software available. Examples of qualitative data analysis software include NVIVO, QDA MINER LITE, or Transana. Examples of Quantitative data analysis software include SPSS, R, or XLSTAT. Choosing the right software will depend on your data analysis requirements and budget.
• The Better Evaluation website provides an overview of how to analyse data
• Qualitative data analysis: a methods sourcebook (2013). 3rd Ed. Miles, Huberman, and Saldana, Sage Publications
• WISE: Web Interface for Statistics Education. This website organises a large amount of statistics resources into one central place. It is also home to a series of interactive, sequenced tutorials on key statistical concepts.