According to CNN's methodology:
A total of 1,003 adults were interviewed by telephone nationwide by live interviewers calling both landline and cell phones. Among the entire sample, 28% described themselves as Democrats, 24% described themselves as Republicans, and 48% described themselves as independents or members of another party.I suspect CNN's under-sampled Republicans. Gallup had 28 percent identifying as both Democrat and Republican just over a week ago, with 42 percent identifying as independent. And the NBC/SurveyMonkey poll for the same period had 33 percent Democrats and 29 percent Republicans, with 36 percent independent (and 1 percent undecided).
Here's CNN's write-up, "Post-convention poll: Clinton retakes lead over Trump."
Compare CNN's 52/43 percent to the L.A. Times Presidential Election Daybreak Poll, which tracks the same panel of 3,000 randomly selected respondents through November. Changes in the poll data reflect changes in the opinions among the respondents, not the opinion of a new sample of voters. As of last night, the Times has Trump up over Clinton, 46.2/42.1 percent.
Thus, according to the Times:
Because of the panel design, “we have the same people every time, so changes in the poll are really people changing their minds,” rather than the result of variations in who answers a particular survey, said Arie Kapteyn, the director of the USC Dornsife center, who pioneered the approach for the 2012 election while at Rand Corp.So, again, as I've been saying, I expect the presidential horse race to even out in the weeks ahead, with Hillary Clinton enjoying a slight advantage in an average of presidential polls. Trump should be considered the underdog, and he's going to get hammered mercilessly by the left-wing press.
The panel design typically shows less volatility than traditional polls. Four years ago, it proved more accurate than most other surveys in forecasting the election result, although “maybe that was beginner’s luck,” Kapteyn said.