A little bit about polling with incumbent presidential race

1. 84

   tsmogposted 3 months ago

   An article by NPR, Early polls don't always foretell the fate of a first-term president. Does anything? (Jan 27, 2024) presents some interesting historical information about polling during the election year of an incumbent president seeking reelection. They discuss different polling questions or types, providing historical evidence with graphics.
   https://www.npr.org/2024/01/27/12255639 … val-rating
   
   One observation is "Early polls may miss incumbent presidents' Election Day strength." (See graphic) They specifically looked at approval ratings. One observation is reaching at or above 50% is key with success of being elected. However, 2 out of 5 times it proved false for incumbents surpassing 50% while Obama in 2011 was at 43% and won.
   
   They also suggested paying close attention to right or wrong direction questions. They proved to predict the Electoral College but not the popular vote. It is advised that attention is paid to the gap between those two.
   
   "It showed the "wrong track" answer was chosen by an average of 70% and "right direction" by just 23% in the five most recent presidential election years when the president (or his party) lost. The gap averaged 46 points in those years (1980, 1992, 2000, 2008, and 2016).
   
   When the numbers for "wrong track" and "right direction" were closer to even, the results were quite different. The two responses were just 7 points apart (49% to 42% on average) in the five election years when the incumbent (or his party) actually won (1984, 1988, 1996, 2004, and 2012)."
   
   With economics, they state and show with a graphic, "Consumer sentiment is not a reliable indicator of Election Day strength." They also state with a graphic, "Unemployment is not a reliable indicator of Election Day strength." Lastly, again, with a supporting graphic, the "Misery index is not a reliable indicator of Election Day strength." The misery index is the combination of inflation with jobless rates.
   
   Questions:
   
   Does this article offer value for observing the polls to arrive through this election year?
   
   Does it offer confidence in polling overall or give cause for doubt?
   
   A bonus to offer. The link next goes to FiveThirtyEight pollster ratings (Updated: Jan 25, 2024).
   
   https://projects.fivethirtyeight.com/pollster-ratings/

   replyreport
   1. 84

      tsmogposted 3 months agoin reply to this

      What does Ipsos say about polling for the rematch of Trump vs. Biden?
      
      The 2024 U.S. presidential election outcome is highly uncertain by Ipsos (Jan 26, 2024)
      https://www.ipsos.com/en-us/2024-us-pre … -uncertain
      
      "At Ipsos, we believe in looking at a wide range of indicators to predict future events. For elections, we frequently employ three independent forecasting models to assess election outcomes: (1) a poll-based model, (2) a heuristics model based on the main issue, and (3) a fundamentals model based on approval ratings and incumbency. We have employed this model-based approach in dozens of elections around the world to great success. Some recent examples include U.S. 2016, U.S. 2020, Brazil 2022, Argentina 2023, and Turkey 2023."
      
      "So, what do they tell us?"
      
      "Today, despite significant challenges, Donald Trump out-performs Biden for the 2024 presidential election. Of our three election models, two of them point to Trump as the favored winner."
      
      
      
      "So, what does this all mean? Should we anoint Trump?  No. While right now, Trump is in a strong position, the nine-month perspective is less than clear."
      
      
      
      Note: FIveThirtyEight pollster rating puts Ipsos at #17 with a 2.8 with 3.0 as the high.

      replyreport