It is a lot easier to do good statistics now than ever before. It’s really important to make sure studies use good methodology.
I’m personally opposed to a lot of meta studies because older studies tend to bin ages and covariates and there is zero logic or evidence to that except it making things seem simpler on paper.
Modern regression spline techniques and knowing to use f tests correctly actually give us much much more reliable models that better use observed data.
As Frank Harrel calls it, dichotomania (arbitrarily binning real values) is a scourge on science.
It is a lot easier to do good statistics now than ever before. It’s really important to make sure studies use good methodology.
I’m personally opposed to a lot of meta studies because older studies tend to bin ages and covariates and there is zero logic or evidence to that except it making things seem simpler on paper.
Modern regression spline techniques and knowing to use f tests correctly actually give us much much more reliable models that better use observed data.
As Frank Harrel calls it, dichotomania (arbitrarily binning real values) is a scourge on science.