About those POLLS...

[ERD50]

The polls on this forum can be fine/fun as a conversation starter, but it just pushes my button when someone tries to come to some conclusion from the responses. They might say something like "Well, it looks like over half of the people on this forum do such-and-such".

I often jump in to say, you just can't draw any conclusions from a self-selected poll. You only know about the people who responded. To represent a cross-section of the population, you need to randomly choose from that population. You can't count on the responders as representing the whole (they might, but it isn't known)

Then I saw this cartoon, that encapsulated it well :

scroll down






















(edit, the apparent original had a caption that adds a LOT - updated)...

-ERD50

 

[ERD50]

I'm not sure where I first saw that, but I did find what appears to be the original source here, or not...

https://sketchplanations.com/sampling-bias

-ERD50

 

[ERD50]

So no one thought this was funny, or ironic?

Or maybe they didn't scroll down? Heck, I'll re-post it to the top than!

-ERD50

 

[ERD50]

And there is an er-org version:

"How many people are tired of polls"

https://www.early-retirement.org/forums/poll.php?do=showresults&pollid=437

-ERD50

 

[Winemaker]

Let's have a poll......

 

[W2R]

So no one thought this was funny, or ironic?

ERD50, I liked it a lot! It really gripes me when people don't *think* about what the results of a poll (or other study) really signify, and what they don't signify.

I didn't respond until now because I just now saw it.

 

[Koolau]

Polls are one of our diversions here on FIRE Forum. I guess we should take a poll on how many people take them seriously (results having anything to do with reality?) My guess - most would say "just a diversion" but I guess we won't know if we don't take that poll.


Of course, since such a poll would be unscientific...