It is so annoying how people in my field persistently seem to think that non-parametric tests are a solution to low cell sizes in studies. I still read this in papers to this day, and even see it in student tests (not as answers, but as part of the test items).

Let's be clear: They are not. They can fill in for missing distributional alignment (mostly residual non-normality). Bootstrapping will also not save your tiny cell sizes, even if you do it.
If you know your cell size is bad, just use the parametric tests (if this is all you know otherwise) and DEFINITELY report your barn-door sized CIs, or incorporate sensible priors and use Bayes (which is what you should be doing in the first place, IMHO).

To me, this is rejection-worthy stuff. It's 2026. Science has progressed past "I wasn't taught that in school". Do better.

#HCI #USEC #stats #academicchatter

✈️ I am traveling to #ndss2026 today to present two of our @SECUSO_Research papers at #USEC and #MADWeb 🤩

The paper „‘I found the text to be encouraging‘ - Evaluating Different Passeord Strength Calculator Designs“ is joint work with Rozalina and Peter. We compared different password strength meter designs and found that fear appeals are not absolutely necessary to motivate users to choose strong passwords.

You can read a version of the paper here: https://s.kit.edu/password-meter

New blogpost up, inspired by @eliocamp on how to make your #rstats projects fully computationally reproducible with docker, pak and renv.

Share, comment, feedback at will :)

https://k4tana.github.io/blog/2026/02/25/Reproducible-Docker-R.html

#statistics #usec #research #academicchatter

#HCI and #USEC folks: we really need a pool of people willing and _able_ to review psychometrics adjacent papers. I'm probably one of the few (I know of 1 other person) people in my field who knows IRT. In broader HCI, I know @mjskay has dabbled with it. Also CTT knowledge is appreciated, there's a few People I know in that regard, but more is always welcome! If I haven't mentioned your name, and you are in the know, please get back to me!

#HCI #USEC #UsableSecurity people: Would you be interested in tutorial papers or conference workshops about scale/item development and validation with CTT and IRT Methods?

Seeing as a lot of you use scales, but very few of you do psychometric work, it might be a beneficial thing to do (if publishable). Thoughts/comments?

If I see you jamming LLMs into your research projects without reason, I will lose respect for you as a scientist. It tells me the following: you either have no interest of your own, or don't care about valuable insights. Both are reasons to suspect you shouldn't be doing what you're doing.

Stop publishing papers because the topic is trending. It's a disservice to science and a waste of money. Have we learnt nothing from COVID, Blockchain, hell even DNA research?

This is also a warning to reviewers: not accepting good work because it's not "in" or part of the hot shit™ cycle is actively kneecapping progress.

Ask yourself: is the cost of using LLMs for my thing worth the benefit? More often than not, the answer is probably no. Have the courage to admit it and justify it to the stupid reviewers that will demand it.

#HCI #hci #usec #academicchatter

Delighted to have had an event for the U.S. Spain Executives Community this evening at the beautiful Barcelona Health Hub at the Recinte Sant Pau Modernista. We had three health tech founders discuss the ins and outs of building a health tech startup in Barcelona.
#usec #barcelona #health #startups

This #USEC article is 20 years old, and anything it mentions is still true for very similar — or even the same — contexts.

Mindblowing stuff.

https://dl.acm.org/doi/pdf/10.1145/1035134.1035159

In usable security science, email encryption usability has evolved into a meme. The way I know this is just by paper titles. The following are titles of actual scientific papers that have been published.

1999: Why Johnny can't encrypt.
2006: Why Johnny still can't encrypt.
2016: Why Johnny still, still can't encrypt.
2018: Johnny can finally encrypt. But does this make him feel more secure?

I suspect Johnny still doesn't feel secure with his new skill.

#USEC #encryption #email #science

EuroUSEC 2024

Only 1 month left to submit to the 2024 European Symposium on Usable Security!

Submit your research and findings to EuroUSEC, a forum for research and discussion on human factors in security and privacy! Discover the perfect synergy of technology, innovation, and tranquillity as EuroUSEC 2024 makes its way to the charming city of Karlstad.

Call for Papers – EuroUSEC 2024 – 30th Sep - 1st Oct 2024, Karlstad, Sweden

https://eurousec24.kau.se/
#cfp #academia #usablesecurity #usec #EuroUSEC

Three things that need to happen in Usable Security & Privacy research:

- Include & allow journals as methods of publication next to conferences (not possible at my uni/many others).
- Abolish conferences or at least hybridise them in a usable and functioning way. It's 2024, we have climate change and what not, mandatory (air) travel should not be the norm. It would also be more inclusive for many!
- Establish Registered Reports as submission formats. They can massively stagger the reviewing time.

#USEC #privacy #research #academicchatter

Tarkov: BEAR oder USEC? Welche Fraktion solltet ihr spielen? - Escape From Tarkov
#EscapefromTarkov #wipe #usec #bear
https://yewtu.be/l22OQB8IugM?t=3