Science changes lives. Good science helps us understand disease, improve treatments, and make better decisions. But science is done by people, and people can make mistakes. Sometimes the biggest problem is not bad intentions. It is a study design that looked fine at first but was never strong enough to answer the question.
That is why I started Vahtian.
Vahtian is a home for open-source tools and practical ideas that help researchers ask a simple question:
Can we trust this result?
Not because someone famous said so.
Not because a model has high accuracy.
Not because the p-value is below 0.05.
Because the evidence has been tested honestly.
Better science starts before the data
Many research problems begin long before the first participant is recruited.
- A study may be too small.
- Measurements may be unreliable.
- Missing data may change the question the study is actually answering.
- A machine-learning model may appear impressive because of data leakage instead of real signal.
These problems are often easier to fix before data are collected than after a paper is published.
That is where Vahtian focuses.
Open tools for researchers
The software published through Vahtian is open source because better research methods should be easy to inspect, improve, and reproduce.
Current projects include:
- RecoverLite — checks whether a planned study design is likely to recover the scientific question it was built to answer, before data collection begins.
- EpiNet — evaluates graph-based analyses with calibration, permutation testing, contestability, and transparent reporting instead of only reporting a single performance score.
- CiteVahti — helps examine whether scientific claims are supported by the evidence presented in a manuscript.
These projects are different tools, but they share the same idea.
Every important scientific claim should come with evidence, uncertainty, provenance, and clear caveats.
Open source and professional support
Not every research team has time to build simulations, audit study designs, or review complex statistical methods.
That is why ForskAI by Vahtian also exists.
ForskAI provides independent research-methods support for universities, research groups, funders, and organizations. Services include study-design review, simulation-based design diagnosis, methodological audits, and statistical consulting.
The open-source projects remain open. Professional services help researchers apply those ideas to their own studies.
What you can expect here
This blog will explain research methods in plain language.
No unnecessary jargon. No hype. Just practical ideas that help researchers design stronger studies, understand their analyses, and avoid common mistakes.
Sometimes we will discuss statistics. Sometimes machine learning. Sometimes study design. Sometimes research integrity.
The goal is always the same:
Help researchers make stronger scientific evidence.
Welcome to Vahtian.