Faulty conclusions happen when you sample from a highly selective group. Two shockers: the sky is blue and the CO Supreme Court once again reflexively rubberstamps Democrat policy.
The statistical problem of sampling from a highly selective group.
I was listening to an audiobook recently when the author delved into the problem of sampling from a highly selective group. It's a fault in statistical reasoning and I thought I would share with you. Knowing where pitfalls in statistics is always a handy thing to have in your toolbox so you can be better prepared to understand statistical results.
Let me give you an example similar to the one the author used.
I attached a collage I made of random basketball team photos off the internet. What do you know about basketball players?
They're almost always tall. This is especially true in the NBA where even the average players are much taller than the average human. Said another way, being a professional basketball player selects for (I'll use PC language here for fun's sake) the "height advantaged".
Now, let's say that I was a Martian who came to earth and wanted to know what it took to become a good pro ball player. I might find correlations between, say, amount of time practicing and NBA stats. I might find correlations between various attitude or personality traits and success.
So, say I find people who are resilient against initial failure and who practice a lot tend to be the best players.
Fair enough. Likely true enough. Did I miss anything though?
As a Martian it didn't occur to me to think through the people OUTSIDE the NBA. I was sampling from a group that was already highly sorted (selected).
If I had sampled everyone in the population and asked what associations there were with being a good basketball player, I'd have picked up on the association between height and success.
By just looking at NBA players, where they're all uniformly mostly tall, the idea that height could enter into things didn't even show up on my radar.
That is the problem of sampling (pulling examples and measuring) from a group that has already self-selected or been sorted.
Good one to watch out for. If you think about it, it shows up often in statistical research and it certainly calls into question the completeness (if not the accuracy) of any results where it's not accounted for.
Ask yourself as you read things: is there any way that the population or group that this author was looking at could have been filtered before the researcher ever got anywhere near them? If there is, ask yourself what could possibly happen if the missing group or data were included.
Wait the Colorado Supreme Court once again rubberstamped the Democrat policy in this state?
Next thing you'll be telling me is that the sky is blue!
Probably shouldn't surprise anyone, but the Colorado State Supreme Court recently said "well, we can't say if it's unconstitutional unless it passes and then we can rule."
Effectively, this ends the court challenges to HH being on the ballot. At this point it's now in voters' hands.
Yet more examples of the Colorado Supreme Court ruling that anything the Democrats in this state do is just fine legally.
Now you know, should you have wondered, why more and more court challenges to Democrat policy are going Federal.
https://www.coloradopolitics.com/courts/colorado-supreme-court-green-lights-property-tax-relief-measure-ballot/article_2a602dd4-3fcb-11ee-b574-372cb435a963.html