To add some context on this Fox poll:
1,002 people total
- 49.5% are men = 466 people
- Of those, about 62% are fathers, according to the Census. = 288 people IF the sample is representative of parenthood (it's not)
= margin of error of 12 points on the margin between Ds and Rs
1,002 people total
- 49.5% are men = 466 people
- Of those, about 62% are fathers, according to the Census. = 288 people IF the sample is representative of parenthood (it's not)
= margin of error of 12 points on the margin between Ds and Rs
https://twitter.com/JamesSurowiecki/status/1558964320628883457
The margin of error roughly doubles again when you looking at differences between margins. So the 28pt "swing" from the may to August poll could be as small as 4pts (likelier) or as big as 52 (hell no).
And again this is IF you assume Census proportions (which is not correct)
And again this is IF you assume Census proportions (which is not correct)
in other words, the Dads Bounce narrative is a bad use of polling
• • •
Missing some Tweet in this thread? You can try to
force a refresh
