*Psychopathic Personality Inventory

Ah I see! Thanks so much!

I guess what I'm saying is it seems like these simulations are perhaps unrealistic?

it's been a while since my grad school measurement course, but I remember learning that the reliability of a variable sets the upper bound of it's (true) correlation with another variable... In such a case wouldn't a delta of .7 with reliability below .7 only be possible due to random chance?

For example, in the rightmost plot: isn't it highly unlikely that the reliability of both predictors is zero or close to zero and yet the correlation between them is .7?

Specifically I'm struggling with the simulations shown in Figure 3. If the reliabilities of both predictors are zero then shouldn't the correlation between them also be zero (on average)? It seems odd to to fix the delta (the correlation between predictors) to different values in such a case?

Giving me GRE flashbacks...

Looks like a cool read thanks for sharing!

Sure! let’s say both variables are normally distributed

Interesting thanks! This is purely hypothetical so I don’t have data to plot. I was trying to simulate data that would behave like this though. I could certainly try adding a discontinuity and see what happens though!

@dingdingpeng.the100.ci

My current thinking is that it would have to be type-1 error because the continuous variable will always (?) have more power than the binarized version of that variable.

Thanks for sharing. it seems like all elected democrats are going to oppose Trump and all elected republicans are going to support Trump… I believe calling helps when there is a niche issue that you think isn’t being recognized. But I’ll try calling and if nothing else will feel like I tried lol

Genuinely curious: will calling elected officials really have any effect on Trump’s agenda (e.g., rescinding of executive orders)? I have trouble believing this, but would be willing to make calls if someone convinced me otherwise…

Great post! Thanks for sharing!

I’m pretty agnostic to this, just curious! I like the idea of empirical syndromes. It seems like the right level of granularity to be clinically useful. Too bad the data don’t agree

thanks for clarifying! It sounds to me like maybe the allure of the syndrome level is that it would provide a direct alternative to DSM style diagnoses. However, if there's no empirical evidence for them it seems odd to keep them in the model...

What is the argument for including them?