Sequential Explorations
Many will use NVivo to explore surface questions and to identify general data patterns.
This tool does enable deeper explorations. (To access the diagram, click the "Details" button at the bottom right of the visual. Then, right-click the image and "Save Image As.")
Beyond Compound Queries and Group queries, which enable depth, researchers can conduct a sequence of actions over data as depicted in the figure.
Remember that when sentiment is coded, the related text for the sentiment analysis is available in each category [whether binary (positive or negative) or in the four-category version (very negative, moderately negative, moderately positive, and very positive)]. Also, when topic modeling occurs, those various topics are collated into sets of text. Also, when word clouds are collected, word sets are a basic output. And when word trees are output, those may be coded as a set. Any file or node, etc., can be queried and processed for further meaning.
Some Answerable Questions
Some answerable questions may be the following types:
Differences between groups in foci:
- What are some differences between groups (by age, by class, by education level, by geography, etc.) in terms of issues of concern based on a particular question or prompt?
- What are some differences between groups (by attitude, by behavior) in terms of issues of concern based on a particular question or prompt?
Plays on sentiment:
- What are particular group's attitudes about a particular issue? What are their respective expressed sentiments around these issues?
- In social imagery, what are some common themes? What are the sentiments expressed around those themes?
- In social videos, what sort of language is used in the transcript versions? What are some common topics? Within those topics, what are common sentiments?
- In microblogging textual messaging (or social network platform poststream textual messaging), what are issues of concern in a particular social media account? What issues engender the most positive sentiment? What issues engender the most negative sentiment? What do respondents seem to feel neutral about?
Manual vs. autocoding:
- What differences are there between a manually coded codebook from bottom-up coding and autocoded topic modeling (also from bottom-up coding)?
- What are differences between manually coded sentiment vs. autocoded sentiment?
Clear Documentation
That said, they also have to keep close record of what steps they took to arrive at their particular insights.
It also helps to have the parameter settings (for each of the data queries) recorded. Parameter settings directly affect the resulting findings and data visualizations.
Discussion of "Sequential Explorations"
Add your voice to this discussion.
Checking your signed in status ...