I pointed out the rhythm of the cycle of Trump support back in mid-June, when he was coming out of the last slump:
At the rate things are going, it seems like there's a month-up and month-down rhythm to the campaign. Peaks occur in the early part of an odd-numbered month, and slumps early in even-numbered months. Fortunately, the election will be held in the early part of an odd-numbered month, favoring the apparent rhythm, although it also predicts a slump during the Republican National Convention in late July. We'll see.
That was based on relative slumps in early February ("Two Corinthians," Iowa caucus), early April (war on women, Wisconsin primary), and early June (La Raza judge, no harm in primaries since he'd already won the nom). Relative peaks were in early March (Super Tuesday I and II) and early May (Indiana primary, locked up the nom). We just saw rising support in the first part of July.
And right on schedule, starting in late July (7/28) and now into early August, he's going through another slump. The two best polls are the USC Dornsife tracking poll (descendant of the RAND poll, a top performer from 2012), and the People's Pundit Daily tracking poll. Regardless of where you think Trump and Clinton are individually, it's clear that their relative standing has narrowed since last week.
Because this fits the earlier prediction about when there would be peaks and slumps, we don't have to attribute this slump to a post-convention bump for Hillary. After her convention, she basically recovered to where she was before, since her convention was so contentious (DNC Wikileaks, DNC Chairwoman resigning, Bernie boo'd by his own delegates for telling them to elect Crooked Hillary, boring speakers, etc.).
We also don't have to attribute it to the Khans -- the narrowing began on 7/28, whereas Khan didn't put on his act until that evening. Trump has more or less not addressed the issue at all, has not "taken the bait," and has been focusing strictly on his main themes since then. And he hasn't made any other gaffes in the past week.
So, while the last time we could have explained the slump by him pushing the La Raza judge case, and the one before that to various "war on women" topics (Michelle Fields, abortion controversy, tweeting ugly pic of Lyin' Ted's psycho wife), this time there is no clear culprit.
"Media firestorm" won't do either, since the media has been constantly railing against him. After the Orlando shooting, they railed on him for renewing talk about the Muslim ban, about gun control, about using the phrase "the gays," Anderson Cooper ambushing FL Attorney General and Trump surrogate Pam Bondi for not supporting gay marriage, and so on and so forth. And yet that didn't stop Trump from climbing out of his slump and heading toward his peak during the Dem convention.
If anything, it looks like the media are reacting to changes in the popular mood. If they sense support wavering, they smell a vulnerable target and pounce. If they sense rising support for a long time, they retreat and stay halfway neutral for awhile -- like Morning Joe covering the GOP convention fairly for a change.
The media, rather than driving public opinion, are more like opportunists chasing after ratings. When Trump is rising, they dial down their attacks. When he's slumping, they unload. They have no spine and no honor, so they aren't about to lead a sustained charge when they're facing increasing resistance from the public. Media treatment is a passive, lagging indicator of what's going on in his popular support levels.
That isn't to say that they treat him fairly at any time -- only that they treat him relatively less biased when his support is rising, and more biased when his support is falling.
So what is driving this cycle? I think it's just a nervous group of voters who are eventually going to vote Trump, but since it is such a risky novelty, get cold feet, then warm back up to him, feel they've gotten over-excited, then cool off again, etc. The USC poll shows women and people aged 65+ as the most variable -- other demographic groups are either more or less constant, or vary by small magnitudes. Women are more risk-averse than men, and old people are more reluctant to embrace radical change.
(Trump consistently leads with ages 65+, but this support rises and falls by large magnitudes, and since the electorate is skewed toward the old, this strongly affects his overall rise and fall.)
I trust that, like the other times, this slump will be followed by another rise. If the rhythm holds, I predict that the VP debate and the 1st and 2nd Pres debates will unfortunately fall in relative slumps (late odd month, early even month) -- again, regardless of how he and Pence actually perform. The nervous parts of the electorate will be going through a jittery phase, no matter what is happening.
Luckily, though, the final debate is toward the end of an even month, and the election itself is in an early odd month -- both of them ending on favorable conditions.
Buckle up -- it is going to be, as always, a bumpy ride toward victory!