By Greta Underhill

In my last post, I outlined my search for a computer-assisted qualitative data analysis software (CAQDAS) program that would fit our Research Unit’s needs. We needed a program that would enable our team to collaborate across operating systems, easily adding in new team members as needed, while providing a user-friendly experience without a high learning curve. We also needed something that would adhere to our institution’s IRB requirements for data security and preferred a program that didn’t require a subscription. However, the programs I examined were either subscription-based, too cumbersome, or did not meet our institution’s IRB requirements for data security. It seemed that there just wasn’t a program out there to suit our team’s needs.

However, after weeks of continued searching, I found a YouTube video entitled “Coding Text Using Microsoft Word” (Harold Peach, 2014). At first, I assumed this would show me how to use Word comments to highlight certain text in a transcript, which is a handy function, but what about collating those codes into a table or Excel file? What about tracking which member of the team codes certain text? I assumed this would be an explanation of manual coding using Word, which works fine for some projects, but not for our team.

Picture of a dummy transcript using Lorem Ipsum placeholder text. Sentences are highlighted in red or blue depending upon the user. Highlighted passages have an associated “comment” where users have written codes.

Fortunately, my assumption was wrong. Dr. Harold Peach, Associate Professor of Education at Georgetown College, had developed a Word Macro to identify and pull all comments from the word document into a table (Peach, n.d.). A macro is “a series of commands and instructions that you group together as a single command to accomplish a task automatically” (Create or Run a Macro – Microsoft Support, n.d.). Once downloaded, the “Extract Comments to New Document” macro opens a template and produces a table of the coded information as shown in the image below. The macro identifies the following properties:

  • Page: the page on which the text can be found
  • Comment scope: the text that was coded
  • Comment text: the text contained in the comment; for the purpose of our projects, the code title
  • Author: which member of the team coded the information
  • Date: the date on which the text was coded

Picture of a table of dummy text that was generated from the “Extract Comments to New Document” Macro. The table features the following columns: Page, Comment Scope, Comment Text, Author, and Date.

You can move the data from the Word table into an Excel sheet where you can sort codes for patterns or frequencies, a function that our team was looking for in a program as shown below:

A picture of the dummy text table in an Excel sheet where codes have been sorted and grouped together by code name to establish frequencies.

This Word Macro was a good fit for our team for many reasons. First, our members could create comments on a Word document, regardless of their operating system. Second, we could continue to house our data on our institution’s servers, ensuring our projects meet strict IRB data security measures. Third, the Word macro allowed for basic coding features (coding multiple passages multiple times, highlighting coded text, etc.) and had a very low learning curve: teaching someone how to use Word Comments. Lastly, our institution provides access to the complete Microsoft Suite so all team members including students that would be working on projects already had access to the Word program. We contacted our IT department to have them verify that the macro was safe and for help downloading the macro.

Testing the Word Macro       

Once installed, I tested out the macro with our undergraduate research assistant on a qualitative project and found it to be intuitive and helpful. We coded independently and met multiple times to discuss our work. Eventually we ran the macro, pulled all comments from our data, and moved the macro tables into Excel where we manually merged our work. Through this process, we found some potential drawbacks that could impact certain teams.

First, researchers can view all previous comments made which might impact how teammates code or how second-cycle coding is performed; other programs let you hide previous codes so researcher can come at the text fresh.

Second, coding across paragraphs can create issues with the resulting table; cells merge in ways that make it difficult to sort and filter if moved to Excel, but a quick cleaning of the data took care of this issue.

Lastly, we manually merged our work, negotiating codes and content, as our codes were inductively generated; researchers working on deductive projects may bypass this negotiation and find the process of merging much faster.

Despite these potential drawbacks, we found this macro sufficient for our project as it was free to use, easy to learn, and a helpful way to organize our data. The following table summarizes the pro and cons of this macro.

Pros and Cons of the “Extract Comments to New Document” Word Macro

Pros

  • Easy to learn and use: simply providing comments in a Word document and running the macro
  • Program tracks team member codes which can be helpful in discussions of analysis
  • Team members can code separately by generating separate Word documents, then merge the documents to consensus code
  • Copying Word table to Excel provides a more nuanced look at the data
  • Program works across operating systems
  • Members can house their data in existing structures, not on cloud infrastructures
  • Macro is free to download

Cons

  • Previous comments are visible through the coding process which might impact other members’ coding or second round coding
  • Coding across paragraph breaks creates cell breaks in the resulting table that can make it hard to sort
  • Team members must manually merge their codes and negotiate code labels, overlapping data, etc.

Scientific work can be enhanced and advanced by the right tools; however, it can be difficult to distinguish which computer-assisted qualitative data analysis software program is right for a team or a project. Any of the programs mentioned in this paper would be good options for individuals who do not need to collaborate or for those who are working with publicly available data that require different data security protocols. However, the Word macro highlighted here is a great option for many research teams. In all, although there are many powerful computer-assisted qualitative data analysis software programs out there, our team found the simplest option was the best option for our projects and our needs.

References 

Create or run a macro—Microsoft Support. (n.d.). Retrieved July 17, 2023, from https://support.microsoft.com/en-us/office/create-or-run-a-macro-c6b99036-905c-49a6-818a-dfb98b7c3c9c

Harold Peach (Director). (2014, June 30). Coding text using Microsoft Word. https://www.youtube.com/watch?v=TbjfpEe4j5Y

Peach, H. (n.d.). Extract comments to new document – Word macros and tips – Work smarter and save time in Word. Retrieved July 17, 2023, from https://www.thedoctools.com/word-macros-tips/word-macros/extract-comments-to-new-document/

by Greta Underhill

Are you interested in qualitative research? Are you currently working on a qualitative project? Some researchers find it helpful to use a computer-assisted qualitative data analysis software (CAQDAS) program to help them organize their data through the analysis process. Although some programs can perform basic categorization for researchers, most software programs simply help researchers to stay organized while they conduct the deep analysis needed to produce scientific work. You may find a good CAQDAS program especially helpful when multiple researchers work with the same data set at different times and in different ways. Choosing the right CAQDAS for your project or team can take some time and research but is well worth the investment. You may need to consider multiple factors before determining a software program such as cost, operating system requirements, data security, and more.

For the Ecampus Research Unit, issues with our existing CAQDAS prompted our team to search for another program that would fit our specific needs: Here’s what we were looking for:

NeedsReasoning
General qualitative analysisWe needed a program for general analysis for multiple types of projects; Other programs are designed for specific forms of analysis such as Leximancer for content analysis
Compatibility across computer operating systems (OS)Our team used both Macs and PCs
Adherence to our institution’s IRB security requirementsLike many others, our institution and our team adhere to strict data security and privacy requirements, necessitating a close look at how a program would manage our data
Basic coding capabilitiesAlthough many programs offer robust coding capabilities, our team needed basic options such as coding one passage multiple times and visually representing coding through highlights
Export of codes into tables or Excel booksThis function is helpful for advanced analysis and reporting themes in multiple file formats for various audiences
A low learning-curveWe regularly bring in temporary team members on various projects for mentorship and research experience, making this a helpful function
A one-time purchaseA one-time purchase was the best fit for managing multiple and temporary team members on various projects

Testing a CAQDAS

I began systematically researching different CAQDAS options for the team. I searched “computer-assisted qualitative data analysis software” and “qualitative data analysis” in Google and Google Scholar. I also consulted various qualitative research textbooks and articles, as well as blogs, personal websites, and social media handles of qualitative researchers to identify software programs. Over the course of several months, I generated a list of programs to examine and test. Several programs were immediately removed from consideration as they are designed for different types of analysis: DiscoverText, Leximancer, MAXQDA, QDA Miner. These programs are powerful, but best suited for specific analysis, such as text mining. With the remaining programs, I signed up for software trials, attended several product demonstrations, participated in training sessions, borrowed training manuals from the library, studied how-to videos online, and contacted other scholars to gather information about the programs. Additionally, I tested whether programs would work across different operating systems. I kept recorded details about each of the programs tested, including how they handled data, the learning curve for each, their data security, whether they worked across operating system, how they would manage the export of codes, and whether they required a one-time or subscription-based payment. I started with three of the most popular programs, NVivo, Dedoose, and ATLAS.ti. The table below summarizes which of these programs fit our criteria.

NVivoDedooseATLAS.ti
General Qualitative Analysis
Cross-OS Collaboration
Data security
Basic coding capabilities
Export codes
Low learning curve
One-time purchase
A table demonstrating whether three programs (NVivo, Dedoose, and ATLAS.ti) meet the team’s requirements. Details of requirements will be discussed in the text of the blog below.

NVivo

I began by evaluating NVivo, a program I had used previously. NVivo is a powerful program that adeptly handled large projects and is relatively easy to learn. The individual license was available for one-time purchase and allowed the user to maintain their data on their own machine or institutional servers. However, it had no capabilities for cross-OS collaboration, even when clients purchased a cloud-based subscription. Our team members could download and begin using the program, but we would not be able to collaborate across operating systems.

Dedoose

I had no prior experience with Dedoose, so I signed up for a trial of the software. I was impressed with the product demonstration, which significantly helped in figuring out how to use the program. This program excelled at data visualization and allowed a research team to blind code the same files for interrater reliability if that suited the project. Additionally, I appreciated the options to view code density (how much of the text was coded) as well as what codes were present across transcripts. I was hopeful this cloud-based program would solve our cross-OS collaboration problem, but it did not pass the test for our institution’s IRB data security requirements because it housed our data on Dedoose servers.

ATLAS.ti

ATLAS.ti was also a new program for me, so I signed up for a trial of this software. It is a well-established program with powerful analysis functions such as helpful hierarchical coding capabilities and institutive links among codes, quotations, and comments. But the cross-OS collaboration, while possible via the web, proved to be cumbersome and this too did not meet the data security threshold for our institution’s IRB. Furthermore, the price point meant we would need to rethink our potential collaborations with other organizational members.

Data Security

Many programs are now cloud-based, which offer powerful analysis options, but unfortunately did not meet our IRB data security requirements. Ultimately, we had to cut Delve, MAXQDA, Taguette, Transana, and webQDA. All of these programs would have been low-learning curve options with basic coding functionality and cross-OS collaboration; however, for our team to collaborate, we would need to purchase a cloud-based subscription, which can quickly become prohibitively expensive, and house our data on company servers, which would not pass our institutional threshold for data security.

Note-taking programs

After testing multiple programs, I started looking beyond just qualitative software programs and into note-taking programs such as DevonThink, Obsidian, Roam Research, and Scrintal. I had hoped these might provide a work around by organizing data on collaborative teams in ways that would facilitate analysis. However, most of them did not have functionalities that could be used for coding or had high learning curves that precluded our team using them.

It seemed like I had exhausted all options and I still did not have a program to bring back to the Research Unit. I had no idea that a low-cost option was just a YouTube video away. Stay tuned for the follow-up post where we dive into the solution that worked best for our team.