Don Lynam

If you are only going to listen to one thing I suggest, make it this. If you were a Gun Club fan in the wayback, try this one on.

Last installment. Dead center of my current sound obsession. Sackcloth and Ashes from 16 Horsepower. Appalachian Goth-punk with a preacher's fire. This band blew me away when I first heard them. If you liked anything I have posted recently, try this.

I know this is just for me at this point, but tonight I return to William Elliott Whitmore and his album Silently, the Mind Breaks. I love his sound. Amazing voice. The sound is percussive and spare. And the vinyl is super cool.

Cool chilling on the elevated dog bed. Sort of.

Tom Waits' Mule Variations definitely lives in the subgenre I am caught up in--primitive, percussive American music with a gospel undertow. Bone Machine lives there too but across town. This is a great album.

Typical end of the evening. Listening to some vinyl and serving as the holder of the evening's particular chew object. Neither Patterson nor I are complaing.

Typical end of the evening. Listening to some vinyl and serving as the holder of the evening's particular chew object. Neither Patterson nor I are complaing.

A little further out from the center of the zip code I have been listening in than his self-titled album, but still very good. I love early Parker Millsap. Y'all might like him too.

A friend introduced me to Closing Time in college and I have been hooked since. He is brilliant.

This is my true jam. Waits is my all-time number one. From Closing Time to Orphans. Well done.

Untitled Arts' Blueberry Pie a la Mode. I am not typically a fan of fruit beers, but OMG.

I will give it a listen.

I dont know the song but the line is great.

Earlier someone asked for best opening lyrics. I went with "The screen door slams, Mary's dress sways." After an evening full of Charlie Parr, I am changing my vote to "take a look at their big house/bigger than they really need." Fight me.

I was super excited to find this on my porch when I got home. I know what I'll be doing tonight.

Don't worry. Here is some music no one cares about as well.
Parker Millsap (self-titled) is revival-tent folk rock—still percussive and Pentecostal, but less primitive. This a great album.

Interrupting the music posts that no one really cares about for a Patterson Hound post that no one really cares about. "Throw the GD ball already, human!"

More of The Bones of JR Jones. I think i like Spirit's Furnace a bit more, but this one is quite good. And the vinyl (media and accouterments) is very nice. As you can see.

Spirit's Furnace by The Bones of JR Jones. Good stuff. Listen to Dry Dirt (Stripped); it is the essence of the sound I am immersed in right now. Pretty vinyl too.

Whitmore’s Animals in the Dark. I didn’t say it yesterday: this album is phenomenal. It’s one of the 30-ish albums in my top ten—top-shelf stuff, start to finish. Soup to nuts, not a weak moment. I’ve been days deep and dozens of listens in, and it just keeps getting better.

Yep. The "Dark Triad" remains stigmatizing, sensationalistic, imprecise, problematic, unscientific, and easily replaceable. It takes me back to the "good" old days of feeblemindness, moral imbecility, and hysterical women.

My recent obsession with Charlie Parr sent me down a rabbit hole where I ran into old friends and made some new ones. For the foreseeable future, it’s primitive, percussive American music with a dark gospel undertow. First up? New friend William Elliott Whitmore’s Animals in the Dark.

Not only is this a very good album (Medicine County by Holly Golightly and the Brokeoffs), it is also a lovely piece of vinyl.

Great work led by Andrew Castillo extending sample size stability analyses to interactions! We've also added a function implementing these analyses to the InteractionPoweR R package: dbaranger.github.io/InteractionP... #rstats

Most studies do not have the samples to test interactions.
They are underpowered.
They yield unstable estimates.
Routine IA testing dumps unreplicable noise into the literature.
We already have too much of that.
So just stop.
8/8

Modal psychology field study: reliabilities = .80, IA r = .05, main effects r = .20, predictor r = .10.
Under these parameters, the POS (with COS = 50% and POS = 80%) is n ≈ 3,800.
7/8

Under more realistic conditions, results are far less encouraging. Required n becomes 150 when COS is 50%. It becomes 245 if POS is 90% and COS is 50%. If we move to a smaller but still too large ES of r = .10, we need 1,000 participants. With reliabilities at .80, we need 1,540 participants.
6/8

Results are encouraging in the most favorable condition—reliability = 1.0, ES for IA 5 to 10 times larger than what can reasonably be expected for an IA (r = .20), the largest COS (100%), the smallest POS (80%). Under these conditions, one only needs 40 participants for stable estimates.
5/8

We also varied: main-effect sizes, predictor intercorrelations, interaction-effect sizes, and reliabilities. Because unreliability biases estimates downward, we placed COS around the observed effect size rather than the population value.
4/8

We varied COS width from 25% to 100% of the true effect. (At 100%, the COS = ES ± 1 ES.)
We varied POS criteria from 80% → 90% → 95%.
Different choices here = different notions of “stable.”
3/8