AI Coding: Code your documents with AI Assist

Introduction

Hello and welcome! I’m Andrias, your host for today’s discussion on AI coding, a newly released feature from the talented developers of MAXQDA. Just two days ago, this feature was unveiled, which means some bugs are possible, although I’ve encountered none so far. This new implementation has piqued my interest regarding its potential applications, and I hope to share that excitement with you.

Introduction to MAXQDA and AI Assist

Before we delve into AI coding specifically, let’s briefly introduce MAXQDA for those of you who may not be familiar. MAXQDA is a well-established qualitative data analysis software that allows researchers to code text effectively. AI Assist is an add-on that integrates a large language model into MAXQDA, enhancing its capacity for qualitative analysis. What’s great about this integration is that it's not mandatory; users can choose to work without it if desired.

The developers of MAXQDA have equipped AI Assist with a flexible model that can be updated regularly based on security and capability evaluations. This ensures that the software remains robust and adaptable.

Goals of the Workshop/Webinar

The goals for today's session are straightforward:

1. Understanding AI Coding: Familiarize yourself with the basic functionality of the AI coding feature.
2. Integrating AI into Research: Explore how AI coding can be effectively combined with traditional research processes.
3. Connecting Features: Highlight how AI coding links seamlessly with existing features within MAXQDA, adding value to your analytical toolkit.

Overview of AI Assist Functions

AI Assist offers several functionalities, including:

• Creating Summaries: The AI model generates text summaries based on various types of input, such as document summaries, code summaries, and segment comments.

• Inspiring New Codes: It can suggest new subcodes based on selected text segments, aiding researchers in developing their coding systems.

• Explaining Terms: AI Assist can provide explanations for unfamiliar terms, which can be a great aid in literature reviews.

• Chatting with Documents or Codes: This allows for interactive inquiries based on the content of a document or code.

Let’s now focus on AI coding, the feature released most recently.

AI Coding Functionality

AI coding allows you to select a document and apply an existing code along with a code definition to that document. While this feature currently only allows for coding one document at a time, it offers speed and efficiency in processing qualitative data.

To demonstrate:

1. Select a document within MAXQDA.
2. Choose an existing code from your code system.
3. Optionally provide a definition for that code.
4. AI coding can then process the document and identify relevant segments based on the definition and focus you provided.

The AI output will not overwrite your existing codes but will create a separate AI-generated code. This allows you to review the AI’s selections against your manual coding, enhancing accuracy.

Integration with Existing Features

AI coding can be seamlessly integrated within the MAXQDA framework:

1. Code System Development: Use AI to suggest new codes and refine your existing codes.
2. Multiple Document Coding: After applying AI coding to multiple documents, you can merge these codes for a comprehensive analysis.
3. Analysis and Validation: The summarizing function lets you compare your manual coding against AI coding to ensure nothing significant has been overlooked.

Conclusion

AI Assist and its coding capabilities present a valuable opportunity to enhance qualitative research. The tight integration with the existing MAXQDA software means that researchers can leverage the best of both technological advancements and traditional methodologies. Using AI as an assistant can lead to greater efficiency and potentially enrich the quality of your analyses.

I hope you find this feature as exciting as I do, and I encourage you to explore its possibilities in your qualitative research.

Keywords

AI coding, MAXQDA, AI Assist, qualitative data analysis, coding system, qualitative research, document summaries, code summaries, interactive inquiries.

FAQ

1. What is MAXQDA?

MAXQDA is a qualitative data analysis software that allows for effective coding of text and analyses of qualitative data.

2. What is AI Assist?

AI Assist is a new feature that integrates a large language model into MAXQDA, enabling users to automate certain tasks in qualitative data analysis.

3. Can AI coding be performed on multiple documents?

Currently, AI coding can only be performed on one document at a time, but you can merge AI codes from multiple documents afterward.

4. How does AI coding affect intercoder reliability?

AI coding can create overlaps in coding segments and produce good reliability, typically around 80% in preliminary testing.

5. Can I define the coding length or segments in language coding?

As of now, it is not possible to specify the length of coding units (e.g., sentence, paragraph) in AI coding.

6. How should I credit AI Assist in my work?

When using AI Assist for coding, be transparent in your methodology by describing its role and the version used in your research documentation.