来自《耶鲁大学开放课程：有关食物的心理学、生物学和政治学 第3集 第3集》

• 
  
  Well we've talked about some of these already,
  我们已经讨论过一些了
• 
  
  but there are a ton of ways
  但当你开始估算人们的饮食时
• 
  
  that inaccuracy can get introduced
  有很多方面
• 
  
  when you start assessing what people are eating.
  会造成误差
• 
  
  And each time one of these errors occurs,
  每当有误差发生
• 
  
  or you get multiple errors that compound each other,
  或多重误差复合叠加
• 
  
  then you've got a lot of potential error
  然后你所得到的信息
• 
  
  introduced in these messages.
  就会包含很多潜在的误差
• 
  
  Okay, first of all,
  好 首先
• 
  
  because everything relies on self-report
  因为任何事都是建立在自觉上
• 
  
  there's no machine, for the most part,
  大多数环节没有机器参与
• 
  
  that can do objective measures of this,
  这样的方式比较主观