I’ll get to the comics soon enough. Interesting me right now is that I made a mistake in my review of the July reading statistics around here. Naturally I want to fix my mistake. But I also thought some about why I thought this an interesting mistake to make. This got me to think a bit about story.
I had made a spreadsheet to work out twelve-month running averages. This for things like the number of page views, number of comments, and number of posts, and all that. Since it was easy to calculate, I also worked out the number of page views per posting. I’m convinced that the number of things I post is the factor I can most control in how well my stuff gets read. And then went on to the number of unique visitors per posting, comments per posting, and number of likes per posting. Fine enough, but I set up the spreadsheet wrong. Instead of dividing the number of unique visitors by the number-of-posts column, it divided by the views-per-post column. And the likes-per-post column divided the number of likes by the number of unique-visitors-per-post. And so on.
I’d like to say I noticed this failed a sanity check. 870 unique visitors for 11 posts, and I claim this to be 7.1 unique visitors per post? Not likely. And then left it in to see if anybody noticed, which of course they did not. No, I didn’t do that; I don’t do that sort of stunt except as a marked joke. Or after warning my class that the story problems might contain unreliable data and they’re expected to ask questions. I did notice the numbers made no sense while writing the statistics-review post for my humor blog, though.
So what do I find interesting about this? Not that I made the mistake. Everyone who works makes mistakes. That I did not notice the mistake is interesting. I can make excuses which of course I find reasonable and justifiable. They all amount to that I chose to do things besides think about what numbers I should expect, and that I did not edit my copy enough before publishing.
Why did nobody notice the mistake? One answer is that nobody read the post, which is plausible enough. WordPress claims the page has gotten 17 views (as of my writing this). My home page, which has the article at its top, has gotten 42 views as well. But a view and perusal are different things. Even if people read my outstanding prose for comprehension, were they reading the numbers? Close enough to notice the claimed numbers didn’t make sense?
My guess is they didn’t. I know when I read for pleasure I tend to accept numbers as things which are present but which don’t need my immediate attention. If the presented argument needs the numbers, I’ll go back and pay attention to whether they’re 7.1 or 79.1. I suspect many people are the same way.
Elmore Leonard famously offered the writing advice to leave out the parts people skip. But people seem to skip these numbers. One might say I skipped them too and I wrote them. This did not make the post unpopular, though. I don’t know why the WordPress readership blog is always a popular post, but it is.
It’s easy to suppose the post would be more popular if it had no numbers. But a readership statistics post without readership statistics? That’s obviously daft. Maybe the box charts and map of countries would be appreciated. Pictures are the other thing besides number of posts that’s within my control and that brings readers.
I think there’s something in the nature of stories going on here. A (nonfiction) narrative builds on facts. If you have none, you may have some fine writing, but you have no story. But a mere fact? We have a word for a bare fact isolated from any narrative, any story about its value: trivia. No one could ever care about the average number of unique viewers per posting around here over the past twelve months. Someone could care about whether this viewers-per-posting is rising or falling, or how fast. The exact numbers, the trivia, are nothing. And we notice this in reading, and accept that we will never care about them. It is the story which uses them that’s of interest, and that people are happy to see. It’s easy, even for a pop mathematics writer, to think that of course numbers are what matters. And they matter, but only a tiny bit. The numbers are there so that the words around them have something to be about. It’s a neat lesson to myself about what mathematics writing means to do.
The correct calculations by the way change the story a little. Not much. This seems weird at first. It supports my contention that the number of page views and unique visitors and comments and likes all scale with the number of posts made, though. A month with twice as many posts probably got about twice as many unique visitors.
I had thought the number of unique visitors per posting rose slightly. Not too much. This is right in kind, but wrong in scale. The twelve-month running average was 60.2 unique visitors per posting and in July there were 79.1. That’s above average enough to matter. I had thought the number of likes per posting went from a twelve-month average of 8.8 down to 6.4. In fact the average was 4.4 and it drifted down to 4.1, still a decline but less sharp of one. One that might not be significant at all. The number of comments per posting I thought had dropped from 3.6 on average to 3.3. In fact the average number of comments per posting had been 1.5, and in July it rose to 1.9. This is the only change in direction of any of these trends. But my suspicion is this is so slight a change that it’s indistinguishable from random fluctuations. Noise, as they say.