John Sides takes issue with my suggestion that "Americans might look at the national jobs numbers to get cues about how they should feel, but their pessimism about the economy is driven by their personal situations, which includes the interactions they have with members of their peer group who may be suffering more than they are."

You can't make that inference from the poll question.  Here is why.  The poll question asks respondents to report on the most important influences on their economic perceptions.  But people are TERRIBLE at stating the actual reasons why they believe what they do.  Here's an old post that cites some relevant research. Here is a post from earlier this week on where people's economic perceptions do come from:

Perceptions of unemployment depend more on the media, but perceptions of gas prices depend more on personal experience. 

In general, my rule of thumb is *never* trust any poll that asks people whether something is important in affecting their vote, views of a political figure, views of the economy, or whatever.  If 60% of the public tells you that Obama's middle name will not affect their vote in 2010, we really can't infer anything from it.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.