Polling shouldn't be about shaping the message. It should be about simply measuring the opinions for better or for worse.
Consider the outcome of this polling effort:
Harvard Millennials Not Convinced
One would think that those young folks immersed in the bath of political correctness and the ideological concepts of the administration would be firmly committed to the belief of the preponderance of support for the incumbent. But, they doubt him. I'd call that good.
They don't, however, see any of the Republican contenders as beating him head-to-head. That might be bad. Is that simply because they are fragmented in their support for a replacement? The question of how the support breaks against a generic opponent apparently was not definitive. It leaves confusion about how Obama loses yet is seen as beating any opponent.
Almost a third of those polled show interest in the Occupy movement. That's good in terms of simple awareness of the world around them. The ivory tower appears to have some windows. But if those interested are supportive of the destruction of the free market and the embracing of redistribution, then that's bad. You can't really determine degree of support since 46% won't express their position. If that is a manifestation of their ignorance, that's bad. If it is a reflection of political pressure to comply and avoid challenging their surroundings, that's bad.
The level of pessimism regarding the direction of the nation is reasonable. Whether good or bad would be determined by how the correction is chosen. If the trend is a return to the values and principles which made the nation great, then clearly good. If the option is revolutionary and manifest through abandonment of our core than certainly bad.
We'll probably all be a lot smarter regarding how things are going to play out next month when some actual votes that mean something are recorded. The handicapping will be over and the actual running will be underway.