#statstab #531 Using Extant Data to Improve Estimation of the Standardized Mean Difference
Thoughts: I wonder if it's a good idea. Any edu/clinical people to chime in? Seems like using historic data.
#statstab #531 Using Extant Data to Improve Estimation of the Standardized Mean Difference
Thoughts: I wonder if it's a good idea. Any edu/clinical people to chime in? Seems like using historic data.
#statstab #506 The VIF Score. What is it Good For? Absolutely Nothing
Thoughts: Maybe we need to think more closely about what these metrics mean.
#diagnostics #assumptions #regression #vif #variance #variableselection #error #multicolinearity
https://journals.sagepub.com/doi/abs/10.1177/10944281231216381
#statstab #505 Beyond Confounders
Thoughts: What makes a good control and a bad control?
#counterfactuals #confounder #DAG #r #modelling #selectionbias #variance #control #causalinference
https://matheusfacure.github.io/python-causality-handbook/07-Beyond-Confounders.html
#statstab #490 Effect size for difference between variances / Levene's test
Thoughts: Always found it odd that we ignore heterogeneity for inference. We treat it like a "error".
#heterogeneity #welch #levene #variance #cvr #standarddeviation
#statstab #489 On the performance of the Neyman Allocation with small pilots
Thoughts: If you know your treatment condition will have larger variance you can optimise your sample size.
#nhst #samplesize #neynan #heterogeneity #welch #variance #pilot #se
https://www.sciencedirect.com/science/article/pii/S0304407624001398
Chaos Edge: Spark Innovation - Reed Hastings and Patrick
Weighting an Average to Minimize Variance
https://www.johndcook.com/blog/2025/11/12/minimum-variance/
#HackerNews #Weighting #Average #Variance #Minimization #Statistics #DataScience
#statstab #393 Statistically Efficient Ways to Quantify Added Predictive Value of New Measurements [actual post]
Thoughts: #392 has the comments, but this is where the magic happens.
#modelselection #modelcomparison #variance #effectsize #tutorial
Researchers have used contorted, inefficient, and arbitrary analyses to demonstrated added value in biomarkers, genes, and new lab measurements. Traditional statistical measures have always been up to the task, and are more powerful and more flexible. Itโs time to revisit them, and to add a few slight twists to make them more helpful.
In case anyone's keeping score, the idea that JFETs have hugely variable properties and that this problem cannot be - or at least has not been - fixed with different design or manufacturing practices, this batch showed it pretty clearly to me.
20 J113s, minus 1 fake, equals 19 devices. I measured the Vgs / pinch-off voltage and the Id / drain current in my test rig. The result?
6 very distinct bins with between 2 and 6 devices in each. Vgs overall varied from (-) 1.30 V to 2.11 V (variance is 38% of the larger value, or 62% of the smaller...), and Id from 5.7 mA to 12.8 mA, a spread of 125% (!) of the smaller value.
No wonder so many JFET circuits just don't work if you don't hand-match the devices.
#variance #variability #characteristics #measure #measurement #match #JFET #tolerance