>>10485644>Are you serious here? The only thing worth a shit for your research question is a longitudinal study. They do not have isolated effects by the nature of this type of research.Yes that's why it's useless to say your research is showing some correlation where for middle part it's a complete flat liner, and some extreme points that number much less than the middle with way worse error bars show negative coefficient.
It's fucking stupid to say "you've got something" when coefficient is completely flat for the most well measured segment of the graph, and then only has a slope because of the less reliable points.
>Now you are simply making stuff up.A point is a point, if you made a point on a graph it's a single measurement. The fact that it's some average over another sample is meaningless. There is a reason you can only make a point out of it, like the fact that it much better shows that all those samples had something in common with each other.
A proper study would just take a representative sample of general population and then plot it. Plotting over averages is yet another thing.
