That was disturbing news to Angelica Jadunandan, who moved to New York from Florida with her two young daughters three years ago. Her older daughter, who is eight, is transgender, and Angelica wanted to move them to a safe blue state—somewhere her family’s medical care would not be called into question. New York is a self-declared “sanctuary state” for trans people. It seemed like a good option.
Oh, I’m not referring to a meta analysis… I actually avoid those because it’s a bit harder to verify all of the sample selection methods. I’m curious about direct, ongoing, large scale studies that may or may not have wrapped up.
You shouldn’t avoid them, because they’re considered the golden standard in terms of evidence. Compiling and analyzing the studies that dozens or hundreds of people have done can give much higher quality evidence than any of those studies alone.
This is true but there’s nuance to it. Often these aggregations have other issues based on their own selection methods.
So, I prefer to look at the direct relationships instead. That’s just my opinion, but I did get it from a person in the field.
I imagine the process differs a lot depending on the subject or focus of the study.
Or I’m just wrong! Oh well 🤷♂️
The issues that individual studies have, on average, tend to be flattened out when aggregated with many other studies. Basing this on what I learned this semester from someone with a PhD in a portion of class specifically devoted to understanding the different types of evidence in scientific studies, and how reliable each type is.
You can always review the methods of a meta study to see if there’s anything that leaps out at you with their selection criteria. But more data is always going to win against less data. Even a study with a p-value of 0.05 can still have results that happened entirely due to chance, but when you add many other studies with the same or similar results then you become much more certain that it wasn’t due to chance.
That makes a lot of sense!