h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

If person A is in 19 out of every 20 public thread, and

If person B posts in maybe 5 or 6 of every 20 public threads, ...

Is person B stalking or harassing person A?

Or is it the nature of a site for people to happen to pass each other in public threads?

If person B's comments are addressed to the world at large, then there is no harm done. But if person B's comments are repeatedly aimed at person A, then they could certainly be accused of harassing them.

it depends on if person B is or not????!!!

Thank you, Gnomon

Maybe person B will now refrain from harrassing A

If person A is in 19 out of 20 public threads, that is, 95% of all public threads,

And person B is in 5 or 6 out of every 20 public threads, that is, between 25% and 30%,

Then it is extremely likely that both persons will be in the same thread, and it is impossible for them not to be in some threads.

Well hey, that first post looks familiar ... When I wrote those exact words some time ago, they vanished twice. (Re-instated both times without debate or appeal too)

Mr Gideon, it would have been polite to acknowledge Nick as the author of your original posting, rather than just borrowing his words.

You could bring this down to a purely statistical calculation: define the hypothesis (eg. "B's posts are more likely to be in the threads in which A has posted"), apply a T-test or something, and come up with 'strong evidence', 'weak evidence' or 'no evidence' to support the hypothesis. That bit's not difficult.

The thing is whether you interpret that hypothesis as meaning 'stalking'. You'd need some qualitative justification, really.

We could also use a Bayesian approach to the problem, Icy. And we probably ought to define "public threads".

I think the Possion distribution is probably the more appropriate.

But also it's all about how people feel. If you are making someone feel stalked it might be worth taking a step back.

Because no one wants to make someone feel bad, right?

Zen, are you joking?

I don't think he was, you know

Yeah, Ok I know the Poisson distribution is totally inappropriate for this situation.

Some folks seeing the same (un-liked) person in two consecutive threads will feel 'haunted and stalked'. The actual mathematics really don't count at that point, it is the emotional response more so than any real or practical one.

Oh, dear.

Dr Zen, Modeling the phenomena as described in the OP as a Poisson event would not be appropriate.

Events that lend themselves to Poisson modeling are rare events, such as component failure or chromosome mutation.

The situation as described in the OP would clearly not be rare.

The actual mathematics really don't count

Don't you ever say that to me again Nick. You *know* how I feel about math.

>> Yeah, Ok I know the Poisson distribution is totally inappropriate for this situation.

Sorry, I started writing my mathy response, went to put the coffee on, fished writing, pressed Post Message then saw this.

Yeah,Possion was the first distribution that came to mind. I don't know why.

Though I don't think you can model based on the fact that the expected response would be for people to post randomly throughout the site. You're more likely to post where you friends are for instance.

In that case you would phrase a hypothesis in terms of B following A where A doesn't follow B. The stalked don't chase the stalkers.

Could I set the record straight that I am not person B.

B