it doesn't take a phd in applied statistics to realize there are 3 issues in data collection today
1 - the sample size is the singlemost relevant predictor of reliability due to tribalism and echo chamber. no one should attempt to extrapolate from any survey with a 500 person sample
2 - the channels each survey uses to collect their sample largely determines the outcome. phone calls skew older, email skews business, tiktok skews under 25 and facebook skews 30-45.
3 - Self-selection bias is also well understood. in our hyper divided world and given trump's ascendance to power again, it is not a surprise democrats would be more motivated to self select and provide survey results when they fell otherwise powerless.
Doubtful there's a conspiracy -- more reasonably its just a function of these 3 issues and either general laziness or lack of the necessary $ and resources to correct the issues on the part of U mich.
Regarding the fact that the economic data is corrupted, if we compare it with the population reality: How many Democrats and how many Republicans are there? This percentage should be replicated or reflected in the sample.
Leaving the bias issue to Bob’s comment, I’d like to comment on the different D inflation expectations vs R. The msm, where most D’s get their news has been fed a steady diet of Tariffs cause inflation. Whereas the alternative media has given a much more nuanced view.
it doesn't take a phd in applied statistics to realize there are 3 issues in data collection today
1 - the sample size is the singlemost relevant predictor of reliability due to tribalism and echo chamber. no one should attempt to extrapolate from any survey with a 500 person sample
2 - the channels each survey uses to collect their sample largely determines the outcome. phone calls skew older, email skews business, tiktok skews under 25 and facebook skews 30-45.
3 - Self-selection bias is also well understood. in our hyper divided world and given trump's ascendance to power again, it is not a surprise democrats would be more motivated to self select and provide survey results when they fell otherwise powerless.
Doubtful there's a conspiracy -- more reasonably its just a function of these 3 issues and either general laziness or lack of the necessary $ and resources to correct the issues on the part of U mich.
Regarding the fact that the economic data is corrupted, if we compare it with the population reality: How many Democrats and how many Republicans are there? This percentage should be replicated or reflected in the sample.
Leaving the bias issue to Bob’s comment, I’d like to comment on the different D inflation expectations vs R. The msm, where most D’s get their news has been fed a steady diet of Tariffs cause inflation. Whereas the alternative media has given a much more nuanced view.