Adam P. Natoli

We are conducting an anonymous online study on personality involving people from all over the world. Participation takes 10 minutes. If interested in participating and you’re over 18 years old, use this link: shsu.co1.qualtrics.com/jfe/form/SV_bs… Please share!!

📢 Call for Papers! Special issue of Assessment on “Artificial Intelligence in Psychological Assessment: Techniques and Applications.” Abstracts due July 21, invited full papers Dec 1. Full call: bit.ly/Assessment_AI_… @APNatoli @lensimms @SagePsychology

[PLEASE SHARE] We are conducting an anonymous online study on personality involving people from 36 different countries and we need your help. Participation only takes about 15 minutes. Use the following link to participate: shsu.co1.qualtrics.com/jfe/form/SV_41…

New article with actual value for clinical applications! doi.org/10.1192/bja.20…

Check out our new paper showing that parent-child relationships are important for immigrant youth who are being bullied at school. doi.org/10.1080/237296…

We are conducting an anonymous online study on personality involving people from all over the world. Participation takes 10 minutes. If interested in participating and you’re over 18 years old, use this link: shsu.co1.qualtrics.com/jfe/form/SV_bs… Please share!!

New open access publication on the implicit measure of personality functioning impairment we developed. Check it out!

Congratulations, @APNatoli 🎉

And the world wants more!! Great collaboration with some exciting results!

Check out our profile for the @NIH AIM-AHEAD grant we received aim-ahead.net/sam-houston-st… @SamHoustonState

@JohnSakaluk Was it Cohen’s (1990) Things I have learned (so far)? tech.me.holycross.edu/files/2015/03/…

Methods Twitter folk: I’m struggling to remember a citation, where the author’s perspective is basically: “don’t use complex analyses (e.g., SEM) over simple ones (e.g., t-tests)”. Ring a bell for anyone?

Dr. Adam Natoli (@APNatoli) and Dr. ABM Islam were recently awarded a grant from the William and Katherine K. Estes fund to host a summer training program at SHSU.

Catch my awkward interview

Check out this fun new article! There are no coincidences: Proposed usefulness of the alternative DSM-5 model for personality disorders in the conceptualization of QAnon-related threats. psycnet.apa.org/doi/10.1037/ta…

Introducing a severity-matching strategy to risk assessment! A severity-matching strategy illustrated using the alternative model for personality disorders in violence risk assessment on college campuses. psycnet.apa.org/record/2023-86… #forensic #psychology #personality

My colleagues, students, and I are doing some cool (and ethical) AI research. @elonmusk Do you have any spare workstations or other hardware that are just gathering dust? Happy to take them off your hands 😁

@WilliamMenton @SFreedenthal I agree with everything William has said. One of the differentials is the researchers’ goal - is the goal to build a model that can be (easily) interpreted and conceptually applied or is the goal to build a model that predicts outcomes well, even if the model is a black box.

@SFreedenthal @APNatoli The included predictors in a stepwise model can be somewhat arbitrary, which challenges inference, but such models can still sometimes produce good, generalizable predictions, even outperforming more theory-based models.

Is anyone surprised that researchers are overfitting their machine learning models? Don’t let fancy “new” methods distract you from the foundational components of good science.

@SFreedenthal It’s possible to replicate stepwise regressions but stepwise regression is almost always misused in our field. That said, it’s less an issue in ML projects when the end goal is a model that can predict things accurately and you don’t care about theory or interpretability

@APNatoli Not surprised! I also gathered that the stepwise regression, rather than hierarchical, was an issue, because stepwise can’t be replicated with a different data set. Is that correct?