#statstab #485 Bayesian ANCOVA and the ATE
Thoughts: Still grappling with the implications of using the causal inference approach to randomized experiments. But it's interesting.
#ATE #causalinference #ancova #ANOVA #rstats #estimand #counterfactuals
#statstab #484 Prediction Interval for a New Response
Thoughts: I think often researchers want to report a PI instead of a CI, at least based on what they claim in the discussion.
#statstab #483 Summary of Mixed Models as HTML Table w/ {sjPlot}
Thoughts: Nobody like formatting tables, especially for complicated models. But you can easy make them with some R code.
#rstats #apa #table #formatting #paper #r
https://strengejacke.github.io/sjPlot/articles/tab_mixed.html
#statstab #482 Introducing Causion: A web app for playing with DAGs
Thoughts: A very cool app. Let's you see exactly what your assumptions and DGP mean for your causal model.
#causal #causalinference #DAG #DAGs #dgp #tutorial #guide #education #pedagogy
#statstab #481 Getting over ANOVA: Estimation graphics for multi-group comparisons
Thoughts: Complex designs are harder to visualise, but with Estimation Statistics you get some perks over simple bar charts.
#design #estimationstatistics #ANOVA
https://www.biorxiv.org/content/10.64898/2026.01.26.701654v1
#statstab #480 Criticism as asynchronous collaboration
Thoughts: Gelman considers this different from a commentary or critique. Interesting. Thoughts?
#metascience #metapsychology #replication #commentary #collaboration #analysis #reanalysis
https://sites.stat.columbia.edu/gelman/research/published/causal_paths_3.pdf
#statstab #479 Different ways of calculating OLS regression coefficients (in R)
Thoughts: The are many ways to skin a variable...
#rstats #regression #modelling #tutorial #r #glm #ols #coding
Many different ways of calculating OLS regression coefficients exist, but some ways are more efficient than others. In this post we discuss some of the most common ways of calculating OLS regression coefficients, and how they relate to each other. Throughout, I assume some knowledge of linear algebra (i.e., the ability to multiply matrices), but other than that, I tried to simplify everything as much as possible.
#statstab #478 Equivalence Tests {marginaleffects}
Thoughts: Often you want to test "no difference" in more complex models than many packages or software permit.
With a few lines of code you can do that for most models.
#Equivalence #noeffect #rstats #TOST #EQ #NHST #hypothesistesting
https://marginaleffects.com/chapters/predictions.html#sec-predictions_visualization
#statstab #477 Don’t calculate post-hoc power using observed estimate of effect size
Thoughts: Good discussion and many useful references. Even big journals print stupid stuff.
#posthoc #power #sensitivity #samplesize #consort #medicine #bias
#statstab #476 Experimental : causal
Thoughts: Randomized experiments are the gold standard for inference for a reason. But they are hard to design.
#design #r #statistics #methods #experiment #tutorial #pedagogy #education #hypothesis #nhst #causal #ancova
https://book.declaredesign.org/library/experimental-causal.html
Dr Mircea Zloteanu ❄️☃️🎄