Here’s a fake story that was shown a number of places on the web during the campaign, claiming that protesters of Donald Trump were being paid. This has been covered so many times by so many fake and satirical sites that it is now an article of faith among Republicans, due to exposure effects.
Here’s a major source of that hoax:
You’ll note the publish date: November 11.
That’s what the site looks like today. But we can see what it looked like previously, courtesy of archive.org’s Wayback Machine.
Here’s what it looked like in March, sporting a publish date of March 24:
Here it is in June, sporting a date of June 16:
And in September it sported a date of September 11:
So it’s safe to conclude that one of the tricks in the fake news toolbelt is creating a feeling of recency through altering dates.
Another note — give the date futzing, I’m not sure we can trust the view counter, but captures from the Wayback do show it ticking up in a reasonable way. If the view counter is accurate (big if) we may also have a ratio of shares to reads.
The page was shared 423,000 times. It was viewed (by people coming through all sources, including but not limited to Facebook) 70,000. If (and again, a big if) we can trust the counter, the maximum click-through rate from Facebook would be 71/423, or 17%. In reality, it would likely be lower than that, as a significant number of people would come through other sources.
To put it another way, at least 83% of people who shared this never looked at it. Note that this is actually more than a recent study that says only 60% of people share without reading on Twitter.
So it’s a big if as to whether we can trust this counter, but if we can, there’s a couple interesting possibilities:
- People share without reading on Facebook more than on Twitter
- More highly viral content has a worse share-to-clickthru ratio
- Explicitly political content has worse share-to-clickthru ratios
None of these are are firm conclusions, incidentally — just ideas I’ll be keeping in mind for the future.