LSE Research: Raising the Quality of Qualitative Analysis

Kavita Abraham, of LSE's Methodology Institute, explains what happens if you get a computer program to analyse qualitative data.

Qualitative data can provide a researcher with a lot of interesting information, but trying to interpret what it means can lead to methodological problems. Unlike a list of numbers, the meaning of a text is not something everyone will necessarily agree on. It's much easier for researcher bias to affect the interpretation, and it's often difficult to explicitly demonstrate how conclusions have been reached.

But what if you could train a machine to do the analysis? In this short film Kavita Abraham of LSE's Methodology Institute explains how she has been using a software called Alceste to analyse the transcripts of interviews with hundreds of people from Angola, Tanzania, and Sierra Leone. The interviews, which explored perceptions of local governance, were conducted by the BBC World Service Trust, a charity that uses the power of the media to reduce poverty and promote human rights. Here Dr Abraham highlights some of the advantages of using a machine to study this kind of qualitative data.

[Levienkhoma]

Amazing software for qualitative research. Researcher spend tones and tones of their time reading through transcripts to identify the codes. This will save us time and energy. Thank you Mi.

[Swasti_Rao]

great presentation kavita. i totally share your view about maintaining the very basic nature of qualitative data as we go broad .

[kiranbains]

So so wrong. There is software. It's called NVIVO. Divorces researcher from interpretation process, which is key to qualitative research. Also ignores methodological pluralism in qualitative research or that different qualitative research projects are underpinned by varied epistemological perspectives. Seems that you want to fit the data to the method, not the method to the data. And ignore tensions between posivitist thinking and post-postivist approaches.