A Nice Debunking

A while back, Suzanne Lucas debunked a claim by one Megan Cornish to the effect that changing her LinkedIn profile to male effectively quadrupled her page views.

I always like a good debunking, but I particularly enjoyed how thoroughly ridiculous she made Cornish's claim look in light of her methodology:

So, it's entirely possible that it's purely a coincidence that her views increased by 400 percent for this week, where LinkedIn thought she was a man.

But she did two other things:

1. Ran her posts through ChatGPT and asked it to change the style to be like a man would write.

2. Asked ChatGPT to make her posts more "agentic."

When trying to isolate a problem, you want to eliminate as many variables as possible to focus on the one thing. By doing three things, it's impossible to tell what the issue is here.

There are many possibilities.

It's the gender swap. LinkedIn denies that its algorithm looks at gender at all, but I've also had LinkedIn employees tell me that adding an external link will not affect views. As a prolific LinkedIn poster, I don't believe that last one for a minute. It's possible that the gender change made a big difference. [bold in original]

Lucas starts with the most charitable possibility first, but she's just being thorough.

It's also the signal for her readers to make some popcorn, because Lucas goes on to consider five other possible explanations Cornish failed to consider or left her results open to by being sloppy.

This was a fun read, and a good blueprint for anyone to remember any time someone makes a broad, fashionable claim like Cornish's.

Just because you're paranoid, doesn't mean "they" are out to get you. Or that they aren't.

-- CAV