 |
|
It seems to me that a better model would be to just take a
Posted on: August 22, 2016 at 09:34:12 CT
JeffB MU
Posts:
72409
Member For:
21.46 yrs
Level:
User
M.O.B. Votes:
0
large random sampling of registered voters from each state, perhaps skewing things a little more towards those who have voted more frequently in the past.
Keeping the same panel would be good, but I'm not sure it's all that much more accurate than a new random sampling each time.
Unless they paid the participants a reasonable amount for participating I think it could skew the results a bit more than if they didn't. If there's a time commitment involved more business people or those working long hours or multiple jobs etc. are going to be more likely to opt out than the retired & the unemployed or underemployed.
No poll is going to be perfect, obviously, but trying to pre-select the percentages of people from pre-set categories seems less likely to give as accurate results as a random sampling would.
Keeping the same panel would be good, but I'm not sure it's all that much more accurate than a new random sampling each time.
Unless they paid the participants a reasonable amount for participating I think it could skew the results a bit more than if they didn't. If there's a time commitment involved more business people or those working long hours or multiple jobs etc. are going to be more likely to opt out than the retired & the unemployed or underemployed.
No poll is going to be perfect, obviously, but trying to pre-select the percentages of people from pre-set categories seems less likely to give as accurate results as a random sampling would.
Please explain why this message is being reported.
MESSAGE THREAD
It seems to me that a better model would be to just take a - JeffB MU - 8/22 09:34:12