Associate Professor @UniVerona, working on #BodyRepresentation, #neuroscience #neuropsychology #bayesian #rstats @Scandle@mastodon.social @scandle.bsky.social
Verona, Italy
Joined May 2008
- Tweets690
- Following1.7K
- Followers977
- Likes1.4K
Michele Scandola retweeted
Negative influence of early social media use on school outcomes dlvr.it/TTr3vL
There does not appear to be a single “NEET personality profile.”
Our new preprint examines how personality traits, psychological resources, and sociodemographic conditions combine in their association with disengagement from employment, education, and training.
Practical implication: one-size-fits-all interventions focused only on “motivation” may be insufficient.
Support may need to strengthen realistic future planning, interpersonal learning, emotional regulation, and the ability to recognize and use available social support.
Important limitations: the sample was relatively small, the design was cross-sectional, and the analyses were exploratory.
The findings do not establish causality, and some unexpected results require replication.
Preprint:
osf.io/preprints/psyarxiv/ky…
#NEET #Psychology #OpenScience
New preprint with Sylia Makhloufi:
Single case analysis in neuropsychology: an Empirical Bayesian perspective
We propose a new methodology based on Bayesian Linear Mixed Models to analyse single-case data in neuropsychology.
doi.org/10.31234/osf.io/c78p…
EBMSC is not free from limitations.
In particular, it may be more difficult for the Bayes Factor to correctly identify the null hypothesis, that is, no difference between the single case and the control group.
We believe that using BMSC and EBMSC together can support a more flexible approach to single-case analysis, while reducing the risk of false inferences.
The article was written in R Markdown for reproducibility and Open Science.
doi.org/10.31234/osf.io/c78p…
In a previous article, @letstido and I introduced the Complex Random Intercepts approach for Linear Mixed Models.
The goal was to simplify random-effects structures without removing key dependencies in the data.
doi.org/10.1177/251524592312…
#BayesianStatistics #LinearMixedModels
Across four simulation studies, we show that Complex Random Intercepts can preserve important dependency structures while improving convergence and reducing computational demands in Bayesian LMMs.
The article also includes a practical workflow and reporting checklist to support transparent and reproducible model specification.
All materials are available at the Zenodo link, including the pre-print and reproducibility files:
doi.org/10.5281/zenodo.21295…