Asst Prof @RiceSocSci. Proud alum @PsychIllinois. Into #iopsych#psychometrics#personality. Here for academic fun and games. Aspiring (b)icon. Views my own.
Our 100-page (!) report (from the 💯 team yooo) also highlighted key limitations with #LLM responses, including social desirability bias and poor item-level fit. These fine details reveal how AI cannot fully replicate the complexity of human responses in psychometrics... yet. 🥸
Can AI replace human respondents in psych research? 🤖
We compared #LLM-generated and human responses using BFI-2 & HEXACO scales. AI mimics broad patterns but misses subtle human nuances at the item level. 📊
Validation is key!
https://t.co/UYlbEpSfjD
#AI#Psychometrics
I'm excited to share that Rice is hiring TWO new faculty members in Cognitive Neuroscience (1 Assistant level, 1 Associate/Full)!
Asst ad: https://t.co/HnfZs6eh3N
Assoc/full ad: https://t.co/x7GD187nz2
Please share, and reach out if you have questions!
On putting the horse (raters and criteria) before the cart (variance components in ratings) | Industrial and Organizational Psychology | Cambridge Core - https://t.co/8D7SWfIYFx
Congratulations to @PanDwww and Hwayeon Meyong!
1st year PhD student showing up at the 8 am panel with 25 expertly crafted slides, 57 backup slides, and print-outs prepared four weeks in advance
vs
Tenured professor giving the keynote with two slides of mood pictures clicked together in the hotel lobby 15 minutes earlier
The IRT Song, another Psychometric Jazz hit dropping in your timeline.
"In the land of stars and stripes, they turned away,
From Rasch's simple truths, they chose a complex way,
Item Response Theory, oh what a tangled dance,
They missed the Rasch measurement advantage,
didn't give it a chance.
To the tune of assumptions, they built a lofty throne,
With parameters three or more, complexity overblown.
Meanwhile, Rasch with its elegance stood quietly alone,
One parameter to rule them all, yet largely unknown.
Oh, in the halls of academia, they debate with such glee,
"Is it IRT or Rasch that truly holds the key?"
But those who've seen the light, they all agree,
The beauty of Rasch is in its simplicity.
So let's raise a glass to Georg Rasch, a visionary so bright,
His model cuts through the noise, brings the data to light.
And to those still lost in the IRT night,
May you find your way by Rasch's clear sight."
Because the Psych Job Wiki is ending, I’ve worked with Amedee Martella to create a new Psych Job Wiki for the 2024-25 academic job cycle! This site has TT jobs, non-TT jobs, and postdoc positions. Feel free to spread the word, use the site, and contribute: https://t.co/F3SchNJzS8
A feature on @KState’s Center for Artificial Intelligence and Date Science (#CAIDS), a consortium of labs led by our colleague @pascalhitzler, PI of @DaSeLab. This article also covers the work of @handekucuk & @tspsyched.
#WCP2024 (personality) and #SIOP2024 (I/O psych) conferences were back-to-back. Research silos are a major pet peeve, so here's a comparison 🧵to facilitate conversations between the two. (1/16)
@David_J_Hughes@cjsotomatic #5. Both conferences had loads on AI. Crowd favorites were Heron, Sylvara, & @tspsyched exploring the ouroboros of job applicants using ChatGPT to fake a chatbot personality inventory and Foster using AI to create narrative feedback reports based on personality scores. (9)
Many people have no idea how these things work.
For example, right now, as a Brazilian, I can't accept invitations to speak at conferences in Europe because my visa to work in the U.S. has expired.
I am not illegal; I have a work contract with the University of California until 2027. Still, I must travel outside the U.S. and get a visa stamp on my passport to allow me to move freely to talk about science.
I have to do this every single year I stay in the U.S. The costs of the visa, traveling, time lost, etc, are all on me. Not every international scientist can handle this.
See what our “always online” group just put out! 🤓This was our international March madness for the win! 🏆Catch me at #siop24 or @guenolen at #imps24 or anyone on the team anywhere to chat more about LLM psychometrics! 🚀
1/In our preprint we proposed PseudoFactorAnalysis (PFA) as methodology based on LLM embedding to assess psychometric properties’ with no data. We compared the results of PFA with those of validation studies. The results were very promising and left space for future development!
A fascinating paradox: People rate AI responses to their emotional problems as more empathic than human responses, but not if they *know* the response comes from AI.
They feel most "heard" by an AI they think is human, least by a human they think is AI.
https://t.co/vfMDEzoPhA
Super fun project on investigating the prediction behavior of LLMs from a psychological perspective🧠
🚨GPT-4 can fall into a cognitive trap and make mistakes even when it has the correct knowledge, like us humans🤯
✅A simple hint helps
Check out our paper for more details👇