Automatic Coding Tools in ATLAS.ti 9 Windows, Mac, & Web
By Dr. Neringa Kalpokas & Ivana Radivojevic
ATLAS.ti 9 offers a whole series of automatic coding tools that make it possible for you to analyze the contents of any of your text documents with the click of a button! This is possible thanks to the integration of power machine learning models that can expertly expand text search queries and automatically identify entities and sentiments. You can thus have ATLAS.ti search through your text data, and the results are saved as coded quotations: ATLAS.ti will select each relevant segment of text and associate a code to each quotation! You can choose which code you want to attach to each quotation, to all quotations, or even use codes suggested by ATLAS.ti. In this case study, we will explore the different “search and code” tools that you can find across the ATLAS.ti platforms. We will illustrate the different applications of the tools using our “Minecraft Evaluation” sample project, which you can download here: https://atlasti.com/learning/sample-projects/
We will go through each of the four main search and code tools, and we will provide an illustrative example of how each could aid analysis in ATLAS.ti 9 Windows, ATLAS.ti 9 Mac, and ATLAS.ti Web. We want to focus on showing how you could apply these tools in your own research, rather than providing step-by-step instructions for each tool. You can see exactly how to use each tool in the manual, video tutorials, or our free webinars. The following table provides an overview of the different search and code tools you can enjoy across the ATLAS.ti platforms.
|Description||Supported languages||ATLAS.ti 9 Windows||ATLAS.ti 9 Mac||ATLAS.ti Web|
|Text search||Automatically search and code for any word(s) and/or phrase(s), including synonyms||All languages (but synonyms are only provided in English)||X||X||X|
|Expert search with Regex||Automatically search and code for any word(s) and/or phrase(s) using GREP||All languages||X||X|
|Named Entity Recognition||Automatically search and code for any people, organizations, locations, or miscellaneous objects (e.g., works of arts, languages, political parties, books, etc.)||English, German, Spanish and Portuguese||X||X|
|Sentiment Analysis||Automatically search and code for positive, negative, and/or neutral emotions||English, German, Spanish and Portuguese||X||X||X|
Are you interested in seeing where and when a particular word or phrase appears in your data? Perhaps you have a specific keyword you want to analyze, or maybe you identified some interesting words from content analyses you conducted by creating word clouds or word lists, and you want to see where exactly in your data these words or phrases are mentioned. The text search tool is your one-stop solution for finding any words or phrases across your entire body of text!
For example, in our examination of how parents evaluate the computer game Minecraft, we noticed that words such as “creative” or “creativity” appeared quite a lot. This seemed like something worth investigating further, since Minecraft is all about building things with seemingly endless possibilities. Since we have a specific concept in which we are interested, we can take advantage of the text search tool to immediately see where and when participants are talking about creativity. Below, we show what these text search results would look like in each of the three main ATLAS.ti platforms (just to show how the different interfaces look – the results are naturally the same since the project is identical across the three platforms!).
Besides just seeing where participants talk about “creativity”, we can also easily add synonymous terms to expand our search and capture anywhere and everywhere participants are talking about the concept of creativity, as opposed to just literally using the word “creative”.
In ATLAS.ti Web, we can repeat this text search in each of our documents of interest. In ATLAS.ti 9 Desktop (Windows and Mac), we can conduct a single text search across multiple documents and/or document groups. Moreover, ATLAS.ti Desktop’s machine learning model also includes the possibility to search for inflected forms of each word (such as plural forms, past/present/etc. tenses, and comparative and superlative forms of adjectives and adverbs).
See the specific steps for using the text search tool here:
ATLAS.ti 9 Windows: https://doc.atlasti.com/ManualWin.v9/SearchAndCode/SearchAndCodeTextSearch.html
Expert Search with Regex
Are you interested in search for specific words or phrases, but you prefer to use GREP (Globally search for a Regular Expression and Print matching lines)? This tool necessitates slightly more advanced knowledge to build a search term, so if you do not want to bother with using GREP, then you can always count on the Text Search tool! If you are familiar with GREP and prefer to conduct your searches using GREP, then Expert Search is the tool for you!
While we could conduct the same search for “creativity” as we did with the text search tool, perhaps now we want to look for specific examples of when people were building or creating something in Minecraft. Using GREP is great for identifying any data that matches some pattern you are interested in, so there is great flexibility in building these search strings and identifying any unit of text you want to see. Here we give a very straightforward example, but you can find more information (and click on the links) in the Regular Expression Search tool in ATLAS.ti Desktop.
See the specific steps for using the expert search tool here:
Named Entity Recognition
Is your research examining what participants say about a person, place, organization, or some other kind of object? If you are interested in analyzing specific entities and you want to quickly and easily see where these are mentioned throughout your text data, then you can count on the named entity recognition tool!
In our analysis of Minecraft, participants also referred to other video games and gaming platforms. If we wanted to examine which video games or platforms were brought up and what participants said about them, the named entity recognition tool could help us examine this in no time at all. Below, we show the results of the named entity recognition analysis. Since we are not interested in people, locations, or organizations participants mentioned, we are filtering our results below to only view those that pertain to the “miscellaneous” category.
In case you notice that any specific find is miscategorized (or you prefer to categorize it differently for your own research needs), you can always change the categorization. You can also choose how you want the automatic code name to appear. When viewing the results, you can apply any code you want (as always), but here ATLAS.ti also suggests codes we could use based on the category and entity. With one click, you can associate these codes to all of the corresponding quotations about that entity, and you have your results saved!
See the specific steps for using the named entity recognition tool here:
ATLAS.ti 9 Windows: https://doc.atlasti.com/ManualWin.v9/SearchAndCode/SearchAndCodeNER.html
Perhaps rather than see what participants are talking about, you want to analyze how they are talking about something. Are they speaking with a positive, negative, or neutral tone¿ What kinds of feelings are participants expressing in your data? If you are interested in analyzing emotions or reactions, then you can make great use of the sentiment analysis tool!
Our Minecraft project was all about understanding how parents evaluate this game, and a natural question we want to answer is, how do parents ultimately feel about the game? With the sentiment analysis tool, we can see all the quotations in our data that convey a positive, negative, or neutral tone. With a click of the mouse, we can also have the different sentiment codes attached to all of their corresponding quotations!
In ATLAS.ti Web, sentiment analysis can only currently be conducted in one document at a time, and you should create your sentiment codes first (i.e., a code for “positive”, “negative”, and “neutral”). We can then immediately see what kinds of feelings are being expressed throughout our data!
See the specific steps for using the sentiment analysis tool here:
Once you have coded your results from any of the search and code tools, you can query and explore your findings just as you would with any of your other coded data. Build connections across the different parts of your project in networks. Export customizable reports of your data from any of the managers or use the query tool to search for quotations that have a specific combination of codes or come from a particular subset of your data. Explore potential emerging patterns across codes with the code co-occurrence table (and visualize the results in Sankey diagrams). Examine and compare code frequencies from the code manager (where you can view bar charts) and the code-document table (where you can view Sankey diagrams). In sum, the diverse search and code tools can serve as powerful complements to your analysis of your rich qualitative data. ATLAS.ti can help you quickly and easily identify words, entities, and sentiments across your text data, and you can then use these findings to build your understanding of your data and develop unique insights from your research.
Kalpokas, N., & Radivojevic, I. (2021). Automatic Coding Tools in ATLAS.ti 9 Windows, Mac, & Web. Retrieved from https://atlasti.com/2021/09/24/automatic-coding-tools-in-atlas-ti-9-windows-mac-web/
About the authors:
Dr. Neringa Kalpokas has dedicated her professional career to qualitative methodology. From her doctoral thesis for which she received the cum laude award in the Complutense University of Madrid to working as a visiting researcher at Harvard University, all of her research projects have been qualitative and carried out with ATLAS.ti. During her 15 years of professional work, she has published numerous articles in a variety of high-impact journals, she has given over 450 trainings, and she has helped over 8,500 people carry out qualitative research. In addition to leading the Europe Team of ATLAS.ti and being the CEO of NkQualitas, she is also a member of the Editorial Advisory Board of the Journal of New Approaches in Educational Research and a professor at the international IE University. Following students’ demand for more rigorous training in qualitative research, she pioneered and taught the qualitative research and ATLAS.ti course at IE University. She continually participates in international conferences to continue sharing knowledge, and she is part of a team of reviewers of articles from high-impact journals. She has repeatedly received awards for excellent teaching in qualitative research. She has also received several research grants from the Ministry of Foreign Affairs and International Cooperation, the Government of Lithuania, and Harvard University. Her latest publications include “Demystifying Qualitative Data Analysis for Novice Qualitative Researchers“, “Teaching qualitative data analysis software online: A comparison of face-to-face and e-learning ATLAS.ti courses“, and “Leading a successful transition to democracy: A qualitative analysis of political leadership in Spain and Lithuania“.
Ivana Radivojevic, a former student of Dr. Neringa Kalpokas’ Qualitative Research course, is passionate about qualitative research and ATLAS.ti. After finishing her training, she was invited to join Neringa’s NkQualitas team and has been participating in numerous qualitative research projects since 2015, resulting in multiple publications in high-impact journals. She is currently the Project Coordinator of ATLAS.ti and is a Senior Professional Trainer. She has given numerous courses, including over 250 webinars, and she has helped over 3,000 people learn to use ATLAS.ti and conduct qualitative research. She continually participates in international conferences to learn and share knowledge with the scientific community. Her latest publications include “Demystifying Qualitative Data Analysis for Novice Qualitative Researchers“, “Teaching qualitative data analysis software online: A comparison of face-to-face and e-learning ATLAS.ti courses“, and “Leading a successful transition to democracy: A qualitative analysis of political leadership in Spain and Lithuania“.