Of the several polls available on the US presidential elections, the RAND poll is probably unique in that it samples the same, consistent group of potential voters each week over the course of the election. By tracking the preferences of the same people, it is possible to obtain a less noisy estimate of how public opinion on each presidential candidate has changed over time in response to their campaign efforts and other events.
One criticism of the RAND poll is that because the same voters are being polled on a weekly basis, this act of constant polling has itself an effect on the results. It may create a psychological incentive for the polled to pay more attention to the current state of the election, and to consider more deeply the merits of each candidate.
In other words, they are more likely to be high-information voters.
This final conclusion seems interesting to me. One often cited flaw with democracy is that of the electorate; we all hope for the wisdom of the crowds, yet we also secretly doubt the intelligence of our compatriots. Low-information voters are more easily swayed by irrelevant variables, and may waver towards poor choices and decisions. At the same time it does not seem right to impose tests of any sort, as that is prone to manipulation and abuse. Of course, what should be done is to raise voter education, or at least attention and interest.
Constant polling seems to force this attentional factor, and may be helpful in reducing voter apathy.