Embarking on your analytic journey with ATLAS.ti – Part 1
Think of your ATLAS.ti project as an excursion into unknown territory. The data material is the terrain that you want to study; the chosen analytic approach is your pathway through it. The tools and functions provided by ATLAS.ti are your equipment to examine what there is to discover. The preparation of the data material is like choosing the right time for the journey. Rain and storm can complicate a planned excursion. This also applies to your project, if, for example, during the transcription, the peculiarities of a computer-aided analysis are not taken into consideration, or if the data file formats are not chosen optimally. A well-designed project set-up is like carefully planning your trip, so you do not make a wrong turn at the first intersection and end up in a dead end. In this article you learn about the various file types ATLAS.ti supports, how to prepare transcripts and how to set up your project. Regarding the latter, it helps to know some technical details. Don’t worry, it is not complicated. Knowing a few things about how ATLAS.ti handles documents will make it easier for you to manage your project(s).
Supported data file formats
In principle, ATLAS.ti supports most textual, graphical and multimedia formats.
|Type of data||Format|
.txt (plain text), .rtf (rich text), .doc(x), .odt (OpenOffice), .htm and .html
In version 8 of ATLAS.ti, documents cannot be edited (!)
|Image and text format|
|Image||.mpg, .gif., .jpeg, .jpg, .png, .tif, and .tiff|
Windows: .aac, .m4a, .mp3, .wav
Mac: aac, .m4a, .mp3, .mp4
The recommended format is: .mp3 files with AAC audio
.3g2, .3gp, .3gp2, .3gpp, .asf, .avi, .m4v, .mov, .mp4, .wmv
Mac: .avi, .m4v, .mov, mp4
The recommended format is: .mp4 files with AAC audio and H.264 video
|Geo data||As data source you can chose among Open Street map, Bing map or Bing Satellite map|
|Survey data (Excel)||Results from an online survey can be imported as case-based documents. It is commonly used for the analysis of open-ended questions. You can however use it for all kinds of data that lends itself to be prepared in this format.|
|Reference Manager||Articles and meta data from reference managers like Endnote, Zotero, Mendeley, Reference Manager, a.o.|
|Evernote||If you collect data in Evernote, you can import them directly from there.|
You can collect data from Twitter, searching for keywords, hashtags, users, etc. ATLAS.ti can collect tweets that are not older than one week.
The tweets from one search will be collected in one document. Thus, with each search, ATLAS.ti adds a new document to your project.
Table 1: Types of data supported by ATLAS.ti
Guidelines for preparing interview transcripts
When you prepare interview transcripts, mark all speakers unambiguously and enter an empty line between each speaker in turn. This increases readability and if you want to use the auto coding feature, this will allow you to code hits within a given speaker unit.
Figure 1: Recommended formatting for an interview transcript
In the sample transcript above, the paragraph marker is visible, showing when the Enter button was pressed. The two speakers in the transcript are marked with unique identifiers:
INT: is used for the interviewer
AL: for Alexander, the interviewee.
Using ‘Interviewer’ or ‘Alexander’ as speaker IDs would be impractical as markers because those words might appear in the text itself. In addition, it is a lot to type and prone to typing errors. The character combinations INT: and AL: are not likely to be found anywhere else. This is essential for using the auto coding tool.
This way of organizing the transcript can be used for any documents that include structuring elements, like dates in historical documents, emails or letters. Although neglecting these best-practice rules will not have a negative effect initially, you may later regret not having used them from the beginning.
Guidelines for focus group transcripts
Everything I wrote above for interview transcripts also applies to focus group transcripts. If you want to compare responses of individuals speakers, each speaker unit needs to be codes. ATLAS.ti can recognize speakers if they have a unique ID like ‘Anne:’ or ‘@Anne:’. Based on these, ATLAS.ti finds all speaker units, and you can automatically code them with both speaker name and other attributes.
ATLAS.ti generates codes from speaker names. Using the @ in front of each speaker name has the advantages that all speaker codes are automatically prefixed with @ and are therefore sorted underneath each other in the code list:
For further information see: https://atlasti.com/2018/08/29/how-to-perform-automatic-focus-group-coding-using-atlas-ti/
I recommend that you name your documents in a way that is useful for the analysis. For instance, include criteria that you already know are important for your analysis like gender, age, profession, location or the date of the interview.
Figure 2: Naming your documents for analytical purposes
Naming your files in this way has the advantage that the documents are already sorted by these criteria. This helps creating documents groups in ATLAS.ti for analytic purposes (see below). In addition, a good analytic name gives valuable information when retrieving data and, overall, adds transparency to your project. If you create reports, the document name is displayed above each quotation.
If you have already created a project before reading this suggestions, you can rename each document in ATLAS.ti: Open the Document Manager, right-click on a document and select Rename.
The aim of this section is to help you understand what is happening when you add documents to a project and to introduce you to a few technical issues that happen behind the scenes. Let’s assume that you have conducted an interview study and have 20 audio-recorded interviews. You transfer the audio files to your computer and begin to transcribe and save the resulting text files somewhere on the computer, using your own system for organizing and storing them. Next, you want to analyze the data with the help of ATLAS.ti. You open ATLAS.ti and begin to add data to your project.
When adding documents to a project, they are copied, converted and stamped with a unique ID and become internal ATLAS.ti files. This means ATLAS.ti no longer needs the original files. However, I recommend that you keep a backup copy of the original source files.
Figure 3: Single user project set-up
The unique ID consists of a combination of letters and numbers and looks something like this: 0e5418ea58a94fd28b7d5bd937f884a2.
It allows ATLAS.ti to recognize each document unambiguously as the name of a document, as not even a combination of name and size can ensure that the content of two documents are indeed the same. This is especially important for merging projects when working in a team. The way you ensure this is for one person to oversee setting up the project and adding documents. Everybody else who will be working on the project needs to wait for that person to share with them a project bundle file.
If you want to view or touch your project as a file, either to share it, to transfer it to another computer or to make a backup copy, you need to export it and save it project bundle file.
Figure 4: Project set-up for teams
If you want to learn more about how to set-up a project for team work, here is a vido that explains every step in detail: https://www.youtube.com/watch?v=u-p0LynYZH4&t=8s
Your ATLAS.ti project is saved within the ATLAS.ti environment. This is a folder on your computer under the AppData Roaming directory on the C drive (Windows) or the application folder on a Mac.
As there are situations where it is not possible to work on the C drive, ATLAS.ti has a function that allows you either to move the library to a different location or to create new libraries. You can create as many new libraries as you want. If you work on your own computer and you have enough space on your C drive, I recommend that you leave everything as is and work at the default location. If you work with ATLAS.ti in a computer lab where you do not have your own user profile, you could create a library on an external disk and take it with you. Another possibility is to export the project after each work session and import it when you continue to work on the project.
Reasons for creating a new library at a location of your choice are:
- You are not allowed to work on the C drive because you work with sensitive data and must work at a specific location on a server.
- You don’t have enough space on your C drive.
- You work in a computer lab, and user data on the C drive are removed every night.
- You want libraries for different purposes. This is interesting for those who work with lots of projects – for example, for teaching purposes. I, for instance, have created a library for all the sample projects that I need for this book. When I teach, I create a library just for the course I am teaching. When the course is finished, I remove the library as I no longer need the projects.
Different from version 7, libraries in version 8 cannot be shared by different users. When you work in a team, each team member works within his or her own library. Projects are shared and united by creating project bundle files and by merging them.
The following videos shows how you can create new libraries and how you can switch between libraries:
One last thing that you need to know is that libraries cannot be created or moved to a cloud sharing service like Dropbox, OneDrive and Google Drive. One purpose of cloud sharing services is that they synchronize data across different devices. This could quickly mess up an ATLAS.ti library and result in incoherent projects. To avoid this, ATLAS.ti prohibits creating or moving libraries to such locations. As ATLAS.ti may not catch all available cloud sharing services, you may be able to outsmart the program. This may, however, have dreadful consequences, as nothing is worse than losing an already coded data set and having to code it all over again. If you want to work in the cloud, use the ATLAS.ti cloud version instead: https://atlasti.com/cloud/
Open ATLAS.ti and create a new project:
The following videos show how you add documents:
All added or linked documents are numbered consecutively, starting with D1, D2, D3 and so on. The assignment of the numbers is determined by its position in the list of documents. The default sort order is by name, i.e. in alphabetical order for each batch of documents that you import.
You can enter a comment for each document. This may not be necessary for all types of projects, but users often do not think of adding information that they already have. My advice is to include all information in your ATLAS.ti project that is relevant for the analysis. When analyzing interview transcripts, researchers often write an interview protocol. But instead of adding it to their ATLAS.ti project, they store the protocols as Word files in some other folder. I recommend copying and pasting the protocols into the comment field of the respective document, so you have all information at one place. The likelihood that you will look at the protocols again is much greater when they become part of your ATLAS.ti project.
When working with newspaper articles or reports, add information about the source, such as a description of the newspaper, its circulation, readership and from where you retrieved the document. If the article or report is available online, you can also add the link to the original source.
Each document that has a comment shows a little yellow Post-it in the document icon.
Figure 5: Commented document in ATLAS.ti Windows
Figure 6: Commented document in ATLAS.ti Mac
After you have added and commented your documents, don’t forget to save the project.
Organizing project documents
When you start a project, you should first consider where and at what level the cases are in your data. Is each document a case that you want to compare to other cases? Or are several documents a case, such as female and male respondents? To ease the handling of the different types of data, they can be organized into document groups. Document groups allow quick access to subsets of your data. They can be used for analytic comparisons in later stages of the analysis. Examples of document groups are the classic socio-demographic variables of gender, age groups, material status, profession, location, etc. For an analysis of newspaper articles, you may want to group by country, circulation and type of newspaper.
Groups can also be useful for administrative purposes in team projects by, for instance, creating a group that holds all documents for coder 1, another group that holds the documents for coder 2 and so on.
It is possible to add each document to more than one group; it is not an exclusive either/or allocation. In a classical interview study, you may want to group a document into groups like gender: female, marital status: single and profession: high-school teacher.
In the following videos you can see how to create groups:
Cases can also be embedded within documents – for example, the different speakers in focus groups. Depending on whether the case is at the document level or within the documents, you must handle it differently. If the cases are inside the documents, you must code them. For focus groups or other structured data, ATLAS.ti can do this for you automatically. See https://atlasti.com/2018/08/29/how-to-perform-automatic-focus-group-coding-using-atlas-ti/ for further detail.
At other times, you can only identify cases within documents if you read through them. The you code those manually. Examples are the various actors that are mentioned in a document, or relationships between actors, locations, contexts, organizations or places. As these codes are different from the thematic codes that are developed when coding data for content, I use a special naming convention for those codes. I will write more about it in a later piece of this series. Here are some first ideas.
I already suggested above to use the prefix @ for speaker codes. For the socio-demographics of those speakers, you may want to use a different prefix like a #. If there is a third type like location, place or context, these can be distinguished by yet another prefix. The aim is to create different sections in your code list, so that it is sorted by those criteria. This makes it easier to find something in list of codes and to filter it.
Figure 7: Naming conventions for codes on the meta level
Computer hard disks can fail; laptops can be stolen. Therefore, it is best to store a copy of your project somewhere else. You do this by exporting a project bundle file. I recommend that you store a copy of your project on a server, in the cloud or on an external drive. Different from libraries, project bundle files can be stored in the cloud.
Project bundle files also need to be used to transfer projects between computers. Thus, you need bundle files when working in a team, or if you yourself work on different computers.
This is how you create project bundle files:
You can rename bundle files. However, renaming the bundle does not automatically change the name of your project. Think of the project bundle file like a box that holds all documents that you added to a project + all your codes, codings, memos, groups, networks, etc. The latter information is stored in an internal project file. Putting a different label on the outside of the box does not change anything that is inside. The ATLAS.ti project file “lives” within the ATLAS.ti environment. You cannot extract it separately from the project bundle file.
Figure 9: Difference between the project bundle and the ATLAS.ti project file
Make it a habit to create a project bundle file after each work session and store it in a safe place. You may keep a few rolling copies of the project bundle and from time to time remove older versions.