Do you highlight all the text you read?

According to this poll, highlighting all the text you read is rather common behavior. Some do it to keep track of their reading position, while others have more varied reasons:

"It feels like a physical connection to the page."

"It doesn't help me read or anything. It's mostly just an unconscious tick."

"I consciously highlight part to reread for comprehension (or to look up a word) then this triggers twitchy highlighting for the rest of the piece."

Just how common is it?

The poll showed very high numbers for this behavior. Maybe too high.

The poll was self-selecting. Someone seeing it and thinking "wow, I'm not alone in doing this" would be much more likely to answer it.

As @gideon_b commented, "your poll is going to have a heavy selection bias".

Experiment

I wanted to run an experiment to find out how common it really is. Not self-reporting in a poll, but based on observation.

Preferably through a one-way mirror in a controlled test environment. But as I wasn't willing to go out and find thousands of real-life test subjects, I thought a little script could do the job just as well.

This had been in the back of my mind for a while, so when an article I wrote became popular, I jumped on the opportunity to run a little test of my own.

I quickly coded up a simple script to tally up how often readers selected text before leaving the page. Nothing fancy, just a small JavaScript snippet to phone home the total tally just after the page is closed.

I added the script to the article and went to sleep. Next morning I woke up to having data from over 10,000 reading sessions.

Assumptions

As most people just leave immediately without reading anything, it would be wrong to include those in the numbers. I decided that 30 seconds would be a fair cutoff before considering someone to be actually reading the article.

How many times do you have to select text before you are a compulsive highlighter? I arbitrarily decided that 5 selections would be the cut-off point.

If you disagree with these assumptions, I made a neat spreadsheet where you can change them to your preferred values and see how the results change.

Results

I would call 11.52% of visitors "compulsive highlighters".

The article was viewed 10815 times, with 4479 sessions lasting at least 30 seconds. Of those, 516 highlighted text at least 5 times.

The top 3 highlighters selected text 288, 250 and 248 times. Ten people highlighted text more than once per second. The average highlighter does one selection every 25 seconds.


If you found this interesting, see my other experiments in the Candy Japan behind the scenes blog.