4. Sentiment Analysis
What is Sentiment Analysis?
In The SAGE Handbook of Social Media Research Methods, Mike Thelwall writes:
"Sentiment analysis is the use of computer programs to estimate some aspect of the sentiment conveyed by a text...Understanding the role of sentiment in communication is an important topic in itself, and being able to identify changes in sentiment over time and differences in sentiment between contexts and objects of discussion is particularly useful for social web investigations" (2016)
Sentiment analysis is thus generally understood as an automated process of coding, whereby a computer interprets textual data and assigns it various sentiment markers (positive/negative most basically, with some programs offering the ability to understand sadness, anger, and other emotions).
Sentiment Analysis in NVivo
In NVivo, your sentiment coding sits along with your nodes and relationships. Sentiments function in much the same way as traditional codes, and you can manually conduct sentiment analysis by selecting text and dragging text over.
However, sentiment analysis in NVivo is a great opportunity to have a quick explore of their "auto-coding" functionality.
Auto Coding in NVivo
In recent editions of NVivo, they've begun to incorporate what they refer to as "Automated Insights". It can be a little rough around the edges, and later in this course we'll be exploring different digital alternatives to some of this functionality
For now, we should understand auto-coding as the ways in which NVivo can automatically interpret and analyse semi-structured data. I've found that this works best for text documents or excel spreadsheets.
Video: Auto Coding
Task: Auto Coding for Sentiment Analysis
Make sure that you have semi-structured text data available in NVivo (The sample projects interviews are in the correct format, or you can download uncoded but still structured versions of the interviews here)
Using the auto code wizard, auto code for sentiment analysis across several files.
Wrapping up
This module has aimed to introduce you to the key functions of NVivo when doing research. It has done this by exploring the main types of coding in NVivo - nodes, cases, and sentiment. Each of these features come with different functionality which may or may not be relevant depending on your methodology and research question.
By the end of this module, you should know how to:
Organise your nodes
Create a case classification with attributes
Code for cases and assign values in a classification
Use auto coding to conduct sentiment analysis
More broadly, I hope this module has introduced key methodological concepts around research organisation and management, as well as the importance of the research question. It has also introduced the method of sentiment analysis.
In the next module, you'll explore the ways in which you can keep track of your research through note-taking, file data management, and project maps!
Thank You!
If you have any feedback on this module, please fill in the feedback form here.
To stay up to date with the NVivo community, you can join the University of Melbourne's NVivo Facebook group here or you can sign up for the monthly community newsletter here.
This module was prepared by Alex Shermon. You can follow him on Twitter here or on LinkedIn here.
If you have any questions about this module, or NVivo more broadly, you can get in touch with Alex via email at alex.shermon@unimelb.edu.au
Last updated