NCA Book Resources

NERT tool

The NCA-Extended Regression Table (NERT) is a tool for combining NCA results with regression-based results in a causal-pluralism multimethod study. It extends the conventional regression table by adding two columns: the necessity effect size and its p-value for those predictors hypothesized and shown to operate as necessary conditions. NERT allows the results of NCA and regression-based methods (e.g., multiple linear regression, SEM) to be compared and interpreted jointly. See corresponding Section 11.3.4 in the book.

NEST tool

The NCA-Extended Solution Table (NEST) is a tool for combining NCA results with QCA results in a causal-pluralism multimethod study. It extends QCA's solution table by adding the required membership score of necessary conditions identified by NCA to each sufficient configuration. This allows the combined interpretation of NCA's necessity results and QCA's sufficiency results. See corresponding Section 11.6.3 in the book.

BIPMA tool

The Bottleneck Importance Performance Map Analysis (BIPMA) is a tool for providing evidence-based practical advice on prioritizing actions and resources, based on the combined results of NCA and SEM. It displays predictor variables as points in a map combining their necessity (NCA) and average effect (SEM) results, helping practitioners identify which conditions to address first to achieve a target outcome. See corresponding Section 11.4.1.7 in the book.

SCoRe checklist

The SCoRe checklist can be used to evaluate the quality of conducting and reporting an NCA study. A minimum SCoRe-score of 60 is required for a publishable study. In addition, the checklist provides targeted recommendations for improvement. See corresponding Section 10.3 in the book.

AI-SCoRe

AI-SCoRe is a series of prompts that can be used to get AI-generated feedback on a document describing an empirical NCA study (e.g., a manuscript). Select an option below.

About AI-SCoRe

Developers: Jon Bokrantz, Gijs van Biezen, Jan Dul
Version 0.0.1  ·  Date: May 30, 2026

What is AI-SCoRe?

AI-SCoRe is an extension of the SCoRe checklist for Strengthening (theoretical rigor), Conducting (data and analysis quality), and Reporting (transparency) an empirical NCA study. AI-SCoRe evaluates an empirical NCA study against published guidelines for good practice using generative AI. It consists of a set of instructions for an AI platform such as Claude or ChatGPT to evaluate an uploaded NCA document (manuscript, paper, publication, etc.). The tool identifies potential points of improvement to inspire critical reflection.

What is the goal of AI-SCoRe?

This tool is intended to help authors improve the quality of their NCA work.

What are the limitations of AI-SCoRe?

The tool is not perfect and will make errors. Users should not treat its output as a final, authoritative verdict; instead they should always critically evaluate the results. The final judgment remains with the author. Since NCA is a relatively new method and is often applied/described incorrectly, an AI platform may have limited or incorrect background knowledge about NCA and may provide incorrect answers to in-depth follow-up questions. Furthermore, note that due to the stochastic and subjective nature of this tool the output may vary even across assessments of the same document, and that a user's own settings in an AI platform may also affect the output. Less advanced models may produce less reliable output.

What can AI-SCoRe evaluate?

Authors may want to upload their document (e.g., manuscript) to an AI platform for getting feedback. Following good practice of GenAI use, uploading others' work without explicit permission of the owner may not be appropriate. Users of AI-SCoRe tool upload documents at their own risk. The AI-SCoRe tool itself can be freely used under MIT licence.

How to use AI-SCoRe for peer review?

The tool can only be used for peer review when consent has been granted by the owner of the document (author, publisher). If you are requested to peer-review a document (e.g., manuscript) and consent was not provided, consider directing the authors towards the interactive checklist or this tool as a means of improving the quality of their work.

What AI platform accepts AI-SCoRe?

AI-SCoRe comes in two versions: regular AI-SCoRe which is a regular prompt for single use in a conversation, and system AI-SCoRe which is a reusable stored system prompt for multiple use. AI-SCoRe consists of a set of instructions that are read before acting. The regular prompt AI-SCoRe is fully portable to any LLM with a prompt field including ChatGPT, Claude, Google Gemini, Microsoft Copilot, Perplexity, etc. The system prompt AI-SCoRe is platform-specific. It is currently available as a Skill for Claude. Although it is formatted to be readily imported as a Claude Skill, we have found that the skill files can also be read and used by other AI platforms. (e.g., Codex of OpenAI). Both versions require that the user uploads the document to be evaluated.

How did we develop AI-SCoRe?

AI-SCoRe was first developed as a system prompt tool using Claude and the "Skills" feature of Anthropic. The SCoRe checklist (https://jandul.github.io/NCA/score-checklist/) and the NCA book by Jan Dul (https://jandul.github.io/nca-book/) were used as primary input and to train the tool. We improved the tool by comparing AI-generated evaluations with evaluations by human NCA experts. The final system prompt version was stripped from Claude-specific instructions to make the regular prompt version.

Use regular AI-SCoRe

The standard prompt to be pasted in the prompt field of an AI platform.

How to use regular AI-SCoRe

Requirements

The regular prompt version can be used on any AI platform with a prompt field.

Step 1 — Download the file

Download the AI-SCoRe.txt file to your computer.

Step 2 — Use AI-SCoRe

Start a new conversation, upload the AI_SCoRe_prompt.txt file (or copy/paste the content of it in the prompt field), and upload the document. Request an evaluation, for example, Use the file "AI_SCoRe.txt" as your instructions and evaluate the document in the document file. Follow the steps and output structure exactly as described in the prompt file.

After the report is provided you can request a pdf file for sharing the report with others, for example, Provide a pdf of the AI-SCoRe report for downloading. Include the book links.

Step 3 — Give feedback to development team

To further develop AI-SCoRe, share your experiences with the developers by answering three questions or by sending an email to the developers [use the Give Feedback link].

Use system AI-SCoRe

The system prompt to be uploaded to the knowledge base of an AI platform.

How to use system AI-SCoRe (in Claude)

Requirements

System prompts are called Skills in Claude and are currently available on all Claude plans.

Step 1 — Download the folder

Download the AI-SCoRe.zip folder to your computer. Do not unpack it.

Step 2 — Enable Code Execution

In Claude: Go in your Profile to Settings → Capabilities and make sure Code execution and file creation is turned on.

Step 3 — Upload the Skill

In Claude: go to Settings → Customize → Skills. Click the "+" button, then "+ Create skill" → Upload a skill, and upload the zip folder. The skill will appear in your Skills list.

Step 4 — Enable the Skill

Click on the score-nca skill and toggle the skill on. You can turn it on or off at any time.

Step 5 — Use the Skill

Start a new conversation and upload your NCA document (e.g., manuscript). Claude will automatically apply the skill when it detects a relevant request — for example, "review my NCA paper" or "apply the SCoRe checklist".

Step 6 — Give feedback to development team

To further develop AI-SCoRe, share your experiences with the developers by answering three questions or by sending an email to the developers [use the Give Feedback link].

Give feedback

Share your experience with AI-SCoRe to help improve the tool. You can skip any question.

1. What was good and should stay?

2. What was not good and should change?

3. What was missing and should be added?

You can also email your feedback directly to gijsvanbiezen@proton.me.

Your email (optional — only if you'd like a reply)

✓ Thank you for your feedback!

Quick Start Guide NCA with Stata

This guide helps to get started with condunting a Necessary Condition Analysis (NCA) data analysis in Stata using the nca package. It covers installing the nca package, running the core analysis functions, and interpreting the results. Please note that access to the Stata software is required to use the nca package. See Appendix B in the book.

Quick Start Guide NCA with R

This guide helps to get started with the NCA package in R (version 5.0.0). It covers installing and loading the package, running the core functions such as nca_analysis and nca_output, and interpreting the results. No prior experience with R is required to follow the guide. See Appendix B in the book.

Demonstration NCA with R

This R script demonstrates how to perform the data analysis of Necessary Condition Analysis (NCA) using the NCA package in R. It walks through a complete example — loading data, running the analysis, and interpreting the output — and can be used as a hands-on starting point for your own NCA applications. For more details, see the corresponding Chapter 9 in the book.