Alex Alekseev

aalexee.comA Taxonomy of AI Experiments | Aleksandr AlekseevAI is everywhere these days. Including economic experiments. But what exactly are AI experiments? What bar does an experiment have to clear to deserve that honorable label? That’s the very question Ch...

21 July

📋 Wrote up the story behind our "A Taxonomy of Al Experiments" paper with Christina Strobel, as well as a quick summary. aalexee.com/post/2026-07... You can check out the full paper at www.sciencedirect.com/science/arti...

sciencedirect.comJournal of Behavioral and Experimental Economics | Artificial Intelligence and Machine Learning in Behavioural Experiments | ScienceDirect.com by ElsevierAs AI technology continues to advance and become more integrated into our daily lives, it is increasingly important to understand how human beings interact with and are affected by AI. In addition, ma...

9 June

🥳Excited to see our SI on "Artificial Intelligence and Machine Learning in Behavioural Experiments" at JBEE completed! @elseviersej.bsky.social www.sciencedirect.com/special-issu... Thanks to the authors, reviewers, my co-editors Monica Capra and Oliver Kirchkamp, and the editor Pablo Brañas Garza

papers.ssrn.comA Taxonomy of AI ExperimentsWe introduce a taxonomy of artificial intelligence (AI) experiments. Our taxonomy produces four types of AI experiments: conceptual AI experiments, stylized AI

16 Jan

🥳 Stoked to share that our paper with Christina Strobel, "A Taxonomy of Al Experiments," has been accepted at the Journal of Behavioral and Experimental Economics! Huge thanks to Christina for collab and to the reviewers for extremely helpful feedback! papers.ssrn.com/sol3/papers.... #EconSky

23 Dec 2025

My Game Christmas Tree. What's on yours?

14 Dec 2025

If you are on the market and it feels tough, know this. You are amazing. You have come this far, be proud of yourself. Rejection hurts, but you are not your rejection. Know your worth, shine your light.

2 Nov 2025

This tip will be helpful for anyone who uses Beamer-made slides (or any PDF slides in fact) with a dual-screen setup, where the slides are projected on a big screen from a laptop in extended display mode. And who cares about their neck. aalexee.com/post/2025-11...

🦆 Last weekend, Regensburg hosted a duck race. 6,000 rubber ducks drifting through a canal. Each duck corresponds to a lottery ticket. The first 100 ducks to finish win prizes. We bought 3 tickets. What are our odds of winning? ➡️ Read more: aalexee.com/post/2025-10...

Clearly, what changes here is the number of non-winning combinations. If you buy m tickets, then this number is "n-m choose k". And then we compute the probability of winning as before, although it does not simplify nicely.

Plugging in the duck race numbers, n=6000, k=100, m=3, we get about a 5% chance of winning. We didn’t win, but on the plus side it was the most rubber ducks I’ve ever seen 🦆

What if you bought more than one ticket, though? What is the probability of having at least one ticket win? It cannot be simply the sum of probabilities of a single ticket winning.

Let’s start slow. Suppose there are four tickets (A,B,C,D). But now assume that you bought tickets A and B. There are still six possible combinations of two tickets: AB, AC, AD, BC, BD, CD. But now five of them lead to a prize, hence the probability of winning is 5⁄6.

Then the number of winning combinations is the difference between the total number of combinations and the number of non-winning combinations. After some algebra, it turns out that the probability of winning is simply k/n.

Applying the formula to the duck race, we get a probability of winning, p = 100⁄6000, about 2%.

It is easy to see the pattern. If there are n tickets in total and k randomly chosen tickets win, the total number of possibilities, which is the number of combinations of k elements from the set of n elements, is simply "n choose k".

But what is the number of winning combinations? It seems that it is easier to find the number of non-winning combinations. This is just the total number of combinations of k elements that do not include our ticket: "n-1 choose k".

Suppose there are 3 tickets (A,B,C). You have A. Two tickets win. All possible pairs are AB, AC, BC. You win in two cases, hence the probability of winning is 2/3.

Now suppose there are 4 tickets (A,B,C,D). Again, you have A, two tickets win. Possible pairs are: AB, AC, AD, BC, BD, CD. You win in three pairs, hence the chance of winning is 1/2

If 1 ticket wins out of 6,000, and you bought 3 tickets, your chance of winning is 3/6000. But what if 100 tickets win? That’s less obvious. Let's build intuition through simple examples

dx.doi.orgThe (Statistical) Power of Incentives | Journal of the Economic Science Association | Cambridge CoreThe (Statistical) Power of Incentives

10 Oct 2025

Thrilled to see my paper "The (Statistical) Power of Incentives" out at the Journal of the Economic Science Association. 🥳 Read it here (open access) 👉 dx.doi.org/10.1017/esa.... #Econsky

19 Sept 2025

How to participate in an ESA meeting Step 1. Attend plenary talks Step 2. Attend sessions Step 3. Attend social events Step 4. Meet new people and catch up with old friends Step 5. Oh no, the conference is over Step 6. Lie down Step 7. Try not to cry Step 8. Cry a lot @ecscienceassoc.bsky.social

19 Sept 2025

Inspired by this amazing graphic that I found at a ramen place in Brno

8 Sept 2025

Big thanks to @fialalenka.bsky.social and @i4replication.bsky.social for the replication games in Brno. It was fun. We loved the songs! 🤘

Lenka Fiala@fialalenka.bsky.social

The Brno Replication Games were a success!🥳 Many thanks to our awesome replicators who spotted a range of coding errors and fixed incorrectly specified models. Many thanks also to @econmuni.bsky.social for having us: you are the reason economists love coming back every year😊

6 Sept 2025

👏 Huge thanks to @milosfisar.bsky.social and the team at @econmuni.bsky.social for a fantastic @ecscienceassoc.bsky.social meeting in Brno #econsky

8 Aug 2025

iOS autocorrects “tex” as “Tex” It autocorrects “xetex” as “XeTeX” It does not autocorrect “latex” Can you guess what it autocorrects as “listed”? 🤔

ssrn.comThe (Statistical) Power of IncentivesI study the optimal design of monetary incentives in experiments where incentives are a treatment variable. I propose a novel framework called the Budget Minimi

30 July 2025

🎉 Excited to share that my paper "The (Statistical) Power of Incentives" has been accepted at the Journal of the Economic Science Association! @ecscienceassoc.bsky.social Thanks to everyone who helped me on this journey. ssrn.com/abstract=408... #EconSky

I have written up a few tips for making writing in TeX a bit more enjoyable. You can read the full post at medium.com/@alexander.g... The main points are below ✍️ #LaTeX #WritingTips

Personalize. Choose a font you love (not just a default), and find a theme that feels right.

Remember, your .tex file is just text. Use any editor you like. Dedicated LaTeX editors are great for typesetting, but you can use other editors. Share your LaTeX writing tips below! 👇

I have written up a few tips for making writing in TeX a bit more enjoyable. You can read the full post at medium.com/@alexander.g... The main points are below ✍️ #LaTeX #WritingTips

First, ditch the split-screen. Close that PDF preview while you're writing. Focus on the words, not the layout.

Next, adjust your editor's line width. A narrower column (80 chars) is way easier to read than a full-screen line.

24 June 2025

🚨 New WP! 📄 “A Taxonomy of AI Experiments” — Aleksandr Alekseev & Christina Strobel (2025) ssrn.com/abstract=529... We introduce a taxonomy to help researchers make sense of existing AI experiments and better design their own. 🧵 Here's a brief overview of our approach:

4/ Natural AI experiments These take place in real-world settings, such as A/B tests run by organizations. AI is used as it would be outside the study. These experiments achieve maximum external validity.

24 June 2025

🔍 Our goal is to help researchers choose the most appropriate type of AI experiment for their question. Let us know what you think.

2/ Stylized AI experiments Here, AI is implemented specifically for a study. These setups allow researchers to explore complex and dynamic human-AI interaction with a lot of control.

24 June 2025

3/ Quasi-natural AI experiments These experiments use existing, real-world AI (like ChatGPT), but in a lab environment. This balances realism with control

1/ Conceptual AI experiments These experiments use AI as a label or framing device, no AI is actually implemented. Common in vignette or survey-based studies. They are easy to run but have low external validity

16 Feb 2024

I wonder if anyone at JPE cares about its readers' eyesight. The text in some graphs is virtually unreadable. It's like 2pt. And yet the font size for panels is always HUGE.

doi.orgThe Moral Machine experiment - NatureResponses from more than two million people to an internet-based survey of attitudes towards moral dilemmas that might be faced by autonomous vehicles shed light on similarities and variations in ethi...

12 Feb 2024

My new favorite quote: "The 95% confidence intervals of the means are omitted owing to their insignificant width, given the sample size (n = 35.2 million)" doi.org/10.1038/s415... #Econsky

7 Feb 2024

It's pretty rare to see a paper on the history of econ thought in Top-5, but Magness, P. W., & Makovi, M. (2023) made it. A fascinating read. doi.org/10.1086/722933

3 Feb 2024

Garcı́a, J. L., Heckman, J. J., & Ronda, V. (2023) use a pretty neat way to visualize ATEs: using stacked bar graphs with the differences between the treatment and control on the top. Seems to be a good solution when you have a bunch of effects to show. doi.org/10.1086/722936

1 Feb 2024

POV: You don't really know a person until you ask them to review a paper

Alex Alekseev reposted

Vincent Arel-Bundock

@vincentab.bsky.social

13 Jan 2024

I'm excited to release the 𝚝𝚒𝚗𝚢𝚝𝚊𝚋𝚕𝚎 for #RStats ! Convert R dataframes to beautiful tables in HTML, LaTeX, PDF, Quarto, Markdown, etc. Easy to learn; minimalist interface; concise syntax; ultra-customizable tables; and zero dependency. vincentarelbundock.github.io/tinytable/

Alex Alekseev reposted

Grant McDermott

@gmcd.bsky.social

26 Jan 2024

Stoked to announce a new release of `plot2`, the powerful and lightweight (0-dep) extension for #rstats base plotting! github.com/grantmcdermo... What's new? A bunch of things, including area plots and facets. Draw plots like the below with simple function calls.

26 Jan 2024

Started reading "The Big Short" by Michael Lewis and it opens with a really cool quote from Tolstoy that I, to my shame, did not know

25 Jan 2024

I've been receiving so much spam from the Journal of This or the Journal of That and the like lately that I am starting to think why don't they just combine their efforts and do the "Journal of Everything Everywhere All at Once"

journals.uchicago.eduUniversity of Chicago Press Journals: Cookie absent

22 Jan 2024

My new favorite quote: "We show that, generically, a mechanism is informationally simple if and only if it satisfies a separability condition that rules out most economically meaningful mechanisms." www.journals.uchicago.edu/doi/10.1086/...

16 Dec 2023

TIL that there is an Online Encyclopedia of Integer Sequences

apoorva lal@handle.invalid

oeis.org/A003024 i'm just reproducing this; the robinson paper is the reference to look at

oeis.orgA003024 - OEIS

11 Dec 2023

Are lab experiments in economics RCTs? I noticed that people tend to refer to RCTs if these are field experiments, but not lab. What do you think? #econsky

8 Dec 2023

What is the single most useful thing (model, empirical method, general academic stuff) that you learned *after* your PhD?