As Mark noted in his Outliers post, White House Press Secretary Robert Gibbs attacked the use of Gallup's daily tracking data today, quipping that "if I was a heart patient and Gallup was my EKG, I'd visit my doctor."
Gallup's Frank Newport responded head-on, defending the value of his own polling data by responding that "...the doctor might ask him what's going on in his life that would cause his EKG to be fluctuating so much," while suggesting that the White House was using Gallup as a convenient scapegoat for President Obama's falling poll ratings.
Newport has a point - our own job approval chart now shows Obama's approval rating less then 2% above his disapproval (although as always the "nose" of our trend line is somewhat unstable). To say that Obama's dropping poll data is "meaningless" is to avoid the substance of the question.
However, Newport's response glosses over the fact that daily tracking data, while not necessarily problematic on its own, can be somewhat erratic and susceptible to sensationalistic interpretation. Take the release Gibbs was asked to respond to as an example: it represented the second time in less than three weeks that Gallup marked drops in Obama's approval with special articles posted to Gallup.com (and sent via blast email to media outlets) separate from their typical daily update, while another special release by Gallup on November 24th featured Obama reaching a new low among whites.
All of this may hold some significance as a trend over time, which is just what daily tracking data allows us to observe. However, sampling error makes it difficult to know when a downturn in approval on any given day is random noise and when it marks a significant trend for the president to worry about. Releases to mark every low the tracker hits, while ignoring commensurate upticks (note that Obama's approval was back up to 50% in today's Gallup Daily poll) result in lopsided news coverage that can weaken the public image of polling itself and make it an easy scapegoat for those, like Gibbs, who may be unhappy with the general trend in the data.
Gibbs' attack on polling data may indeed have been a convenient excuse for Obama's dropping poll numbers, but Gallup sets itself up for criticism by overemphasizing daily noise and provides Gibbs with an opportunity to avoid dealing with the substance of the issue and assail sensationalistic media coverage instead.