It’s all about equal time, isn’t it? And that time has come, time to restitute our crack, but Trump-o-centric election reportage, by spreading out some space for the tweets of the Democratic all-but nominee, Hillary Clinton.
I’ve returned to the scene of all those earlier downloads, twdocs.com, once again contributing my $7.80 to the greater journalistic good and coming away with 2384 tweets by the candidate – more-or-less, a qualification that needs to be explained.
But first consider the above number – no, not the $7.80, but the tweet count. Twdocs will recall up to the last 3200 tweets meeting a searcher’s criterion, in my case consisting simply of tweets issuing from the @HilllaryClinton account. That the result fell notably short of the available max suggests Clinton simply isn’t tweeting as much as her opposite number, whose 3111 tweets we counted in previous posts gave me something closer to my money’s worth (note: when calling up the Clinton tweets I requested that replies be held back from the download, a stricture I didn’t introduce into the Trump requisition. I had assumed that the replies had been composed by others, but now I’m not sure. In any event, I doubt these contribute markedly to Clinton’s tweet activities).
You may recall that the voluble Mr. Trump’s monthly tweet breakout went this way (remember his tally took us through May 11):
But the Clinton count, through June 7, reads:
Nevertheless, we see that in recent months the numbers begin to compare, with Trump’s late-2015 barrage accounting for the differential.
But there’s another communicative difference between the candidates, an attributional one. Clinton’s Twitter description, clearly phrased to have us know that she is almost just like you and me, more or less, reads: “Wife, mom, grandma, women+kids advocate, FLOTUS [that’s First Lady of the United States], Senator, SecState, hair icon, pantsuit aficionado, 2016 presidential candidate. Tweets from Hillary signed -H”.
OK, she has two points in common with Trump – presidential candidate and hair icon – but I’m really interested in that last sentence. Tweets bearing Clinton’s personal authorship receive her identifying -H (Barack Obama’s self-disseminated tweets carry the -bo signifier), begging the obvious question about the proportion of Clinton tweets that are incontestably hers. In fact we’ve explored the means for learning such things in the Trump posts, and so for a first pass at the problem I’d name Clinton’s Text field (the one bearing the actual tweet contents) tweets and enter, somewhere:
=COUNTIF(tweets,”*”&”-H”&”*”)
That expression returns a 167, or 7.01% of all the tweets my download delivers. That’s noteworthy, at least for starter, and we can continue as per the Trump analysis by proposing a set of search terms we might expect to find among the Clinton tweets. Here’s what I tried, again for starters, again each guided by this formula:
=COUNTIF(tweets,”*”&[cell address of search term]&”*”)/counta(tweets)
Nothing electrifying in there, but the Clinton tweets, be they self or other-authored, cultivate a keenly non-incendiary persona, with not a single “lyin’”, “goofy”, or “crooked” rough-housing their way into the text (though 9 tweets admit the slightly more Presidential “lying”). The count for Bill, of course, may not be what it seems, at least not if we’re canvassing references to the candidate’s husband; some of these could espouse legislative matters instead, after all, but absent a close content analysis that ambiguity comes with the package.
And indeed, a bit of review persuaded me that the 167 tweets I’ve been laying directly at the word processor of the candidate herself is likewise overstated. A closer look at that yield turned up six tweets exhibiting rather a -Hillary tag, all of which punctuate verbatim quotes from Clinton (three each tweeted in March and April) that were obviously uttered some time earlier. Thus the unmediated Clinton tweet total comes to 161, or 6.75% of them all. (A search for -H[space] wouldn’t palliate the problem either, because very many of the -H closings are the very last entry in the tweet, with no space following.)
I next blocked out the same Month/Search term matrix I imposed on the Trump data (detailed here), and started with these terms:
(Recall here the six -H tweets that aren’t Clinton-authored.) The large Hillary accumulation, amounting to 27.98% of all tweets, probably won’t electrify either, though she and her ghost-tweeters seem to have been exercising more restraint of late; but what’s authentically striking is the spike in tweets calling on – and calling out – Donald Trump (the June count projects to 171), and the tiny tweet apportionment devoted to Clinton’s Democratic opponent, Bernie Sanders. Perhaps the distributions reflect Clinton’s desire to renounce intra-party antagonisms that might come to bite her later, along with a companion sense that in Trump she has sighted her true competitor, one who in her judgment is so fault-ridden that the targeting comes easy.
Note as well the downturn in tweets stemming from the candidate herself, even given the overestimates considered above (the June extrapolation is 13). The samples are small, but perhaps Clinton’s decided that, now that the nomination is all but hers, she has better things to do.
For example, letting me know how to pronounce FLOTUS.
spreadsheetjournalism 
Leave a Reply