kondeor.ai · Blog

Blog

15 November 2026

Researchers Are Learning Differently, Not Less: AI in Data Analysis

Are researchers learning less when AI takes over parts of data analysis? Drawing on Harald Dragan's "wheel that never stops spinning," a post on the new kondeor.ai blog identifies the areas where researchers must remain competent themselves: the research question, the measurement levels of variables, the hypotheses, and a whole range of points related to data analysis. My conclusion from 30 years of practice and teaching: the approach to learning and conducting data analysis is shifting from "How do I have to do it?" to "What can I do (at all)?"

Read more →
Students 26 October 2026

Are 60 Questionnaires Enough? Calculating the Sample Size for Your Thesis

First published at Statistik + Beratung / Daniela Keller

How many participants does a study need? There's no one-size-fits-all answer, but it can be calculated in advance. In this guest post on Daniela Keller's blog, I use a bachelor's thesis as an example to show how a power analysis justifies the required sample size. Comparing two groups with a medium effect size requires 64 people per group, or 128 in total. With the originally planned 60 questionnaires, the power would be only around 48 percent. The post also explains the key parameters — effect size, significance level, and power — along with precision planning for proportions and means, and common pitfalls. These include planning for dropouts and handling multiple hypotheses. The sample size calculator in kondeor.ai takes care of the calculations for you.

Read the original post →