She is probably assuming he’s normal, though.
I’m not responsible for this either
She’s actually arguing against herself there too…I’ve always been told that outliers should simply be ignored…
I didn’t get the tractor one at all, though.
Richard Carter on the tractor post provided a link that may help.
Hmmm. I have a question about the independence of the variables in Example 1. Do the boyfriend’s various dates know one another, and so might be manipulating the results? I bet they do, even if he thinks they don’t. In which case the assumption of normality would be violated and you’d have to subject all data to an arcsine transformation. Or uses some nonparametric test.
And the tractor one is priceless. I’ve seen examples of the second tractor, especially when revving up to escape from muddy fields.
Isn’t that stranger tractor just the ordinary tractor with its wheels going, like, really fast, man?
Ah, just noticed the link, and think I can now safely answer myself.
Thanks for the link. That tractor is pretty funny now that I get it!
I’m glad you’ve all enjoyed them!
I’ve edited my text a bit, to make it a bit pithier. So, no. You’re not going mad.
Bahahahaha, that’s so cool.
But yea, like Christian, I was taught to write off the outliers as random chance…
hehe, Henry has some good pointers on what if the dates in the box and whiskers plot were manipulating the results. it’s probably what skewed them from the outlier. But she convinced him it’s irrefutable, and I guess that’s what counts 🙂
The distribution of time spent together is probably quasi Poisson (it is going to be right skewed as negative values aren’t possible and it is sort of count data) so a transformation is required. Afterwards she probably wont look like an outlier.
Ah, somebody must be leptokurtic. There’s an old cartoon about that too, but I can’t find it at the moment (it’s by George Box, and uses a kangeroo and platypus to demonstrate kurtosis).