4 ms·
Meta question: is there a name the the psychological impact of putting an arrow on a random line where it really does make it look like that's an inflection poi
by JumpCrisscross 2y ago
Meta question: is there a name the the psychological impact of putting an arrow on a random line where it really does make it look like that's an inflection point? Does this generalise beyond charts, e.g. to a narrative techique?
- gpderetta 2y ago> Does this generalise beyond charts, e.g. to a narrative techique? "Scare quotes"? Random Italic for emphasis?
- lokimedes 2y agoFraming bias or perhaps attentional bias.
- hn_throwaway_99 2y agoIf your implication is that the red arrow is just random, are we not looking at the same graphs? I'll grant that it looks to be cherry picking on maybe a couple of the graphs, but clearly appears that something significant and non-random happened in the 1970-1972 period.
- JumpCrisscross 2y ago> clearly appears that something significant and non-random happened in the 1970-1972 period How are you cutting it off that cleanly? How about this: pick a chart and I'll find the data. We can then look at the numbers and see if something significant happened in '71.
- hn_throwaway_99 2y agoSure, how about the first 2 graphs on the linked site?
- bryanlarsen 2y agoIn both cases the point of divergence appears to be ~1975.
- hn_throwaway_99 2y agoAre we looking at the same graphs? To me, in both cases, there seems to be a clear change in the wage growth line right at 1971-72.
- bryanlarsen 2y agoThe only portion of the graph with divergence around 1971 in the first two graphs is the GDP line. The first graph is even worse -- the divergence is higher in the late 60's than it is in 1971.
- jcranmer 2y agoSomething like the Texas sharpshooter fallacy? (The idea being that the Texas sharpshooter proves he's good by shooting at the barn a couple of times, and then paints the target only after he knows where he hit).
- yobbo 2y agoAny graph of an exponential will look mostly the same regardless where you start on the horizontal axis, since it is scale invariant. However, what is seen in these graphs is several data series diverging after 1971-1981. The only suspicious thing is pinning it on 1971 specifically.
- JumpCrisscross 2y ago> what is seen in these graphs is several data series diverging after 1971-1981 Note that all the divergence charts cut off at WWII. (The lines that don't show regressions to a mean.) The charts highlight a problem. They just finger the wrong cause.