Hi all,
I have recently conducted a regression analysis using the Hayes PROCESS macro. For those of you who are aware, the output of this macro spits out data points that can be plotted to create a line graph. The issue is, when the lines were plotted they turned out bent/curved rather than as straight lines. To note, a regression imputation was done to handle the missing data, although I am not sure if this would be the issue here. I have included a photo of the graph that was created. Any thoughts as to why this might be happening would be greatly appreciated.
They are not! Both are linear - it looks like a bend, but there actually isn't! :-)
Zoom in and see for yourself!
I've looked carefully, and sure enough, the Dads' curve becomes steeper, after the intersection with the Moms' curve. Although yes, the Moms' curve is a straight line.
My question is, though, why should we assume that straight lines are correct? I admit that the sharp bend looks suspicious, perhaps an artifact of the algorithm used. But if these curves both became steeper, as the symptom severity increased, would that be surprising?
I don't know what this GARS-3 scale, or the PROCESS macro scale for stress, look like. If anything is surprising to me, it's how these two, apparently unrelated scales, create lines that are straight, or nearly so. What are the odds??
You're right, but it took a real blow-up to notice. What is the message from those two plots?
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs