Research Practice - csen2025.org

Visualizing the multiple comparisons problem.

Test Twenty Hypotheses and One Will Look Significant

I remember sitting in a windowless lab during my PhD, staring at a cluster of p-values that looked absolutely beautiful—until I actually looked at them. I had run twenty different variations of a distributed consensus test, and by some miracle…

Visualizing confidence intervals in reporting data.

A Range Communicates Honesty That a Point Estimate Hides

I remember sitting in a windowless seminar room during my PhD, watching a senior researcher present a single, glowing point estimate as if it were an immutable law of nature. He spoke with such absolute certainty that the room practically…

Effect size and practical relevance comparison.

Statistically Detectable and Practically Useless Coexist Easily

I remember sitting in a windowless seminar room during my PhD, watching a postdoc present a series of charts where every single result was marked with a pristine, glowing asterisk. He was celebrating a “highly significant” breakthrough, but when I…

Understanding p values and their misuse.

A P Value Does Not Tell You the Probability You Are Right

I remember sitting in a windowless seminar room during my second year of my PhD, watching a senior researcher present a dataset that was clearly noisy, yet he was treating a tiny significance threshold like it was a divine revelation.…

Statistical significance explained: unlikely by chance.

Significant Means Unlikely by Chance, Not Important

I remember sitting in a windowless seminar room during my PhD, watching a senior researcher present a series of charts where every single result was “significant,” yet the actual effect sizes were so microscopic they wouldn’t have moved the needle…

Choosing evaluation metrics for optimization success.

The Metric You Optimise Becomes the Thing You Get

I remember sitting in a windowless lab during my postdoc, staring at a loss curve that looked absolutely beautiful, while the actual system was collapsing in real-time. I had spent three weeks optimizing for a metric that looked impressive on…

Importance of control groups and baselines.

Dr. Ingrid Falk-Weller
February 24, 2026
(0)
Research Practice

A Weak Baseline Makes Any Method Look Strong

I remember sitting in a windowless lab during my postdoc, staring at a cluster of results that looked absolutely revolutionary—until I realized we hadn’t actually accounted for the thermal drift in the hardware. We were celebrating a “breakthrough” in latency…

Dr. Ingrid Falk-Weller
February 17, 2026
(0)
Research Practice

Change One Thing or Learn Nothing

I remember sitting in a windowless lab during my postdoc, staring at a cluster of results that looked absolutely beautiful—too beautiful. I had spent three months optimizing a distributed consensus protocol, only to realize a week later that my “breakthrough”…

The Gap Is Usually Smaller Than You First Think

I spent three years in academia chasing “novelty” like it was some mystical, shimmering prize, only to realize I was just rearranging deck chairs on a sinking ship. I used to think that finding the gap in prior work meant…

Applying various literature review methods.

Reading Widely Before Deciding What Is New

I remember sitting in a windowless university basement three years ago, surrounded by half-empty coffee cups and a stack of printed papers that felt more like a mountain than a foundation. I was trying to force myself to follow a…