Polling Measurement Errors: Examples

1. Do Americans believe scientific facts? Or do the tests just measure political affiliation or religiosity? Atlantic

2. Effect of timing: If we’d based our conclusion on the 2016 poll results we might have concluded that we are measuring attitudes toward executive power: I.e., Republicans are more concerned about it than Democrats. Once Trump was elected and Obama was out of power we see a flip. This indicates that at least part of what we are measuring is opposition to the other team–not attitude toward executive power.



3. thumbnail_Untitled-trump.jpg