Imputing new worlds with multiversal appeal

Monthly Archives: August 2016

Sometimes we keep them in,

Sometimes we take them out

It all depends,

On what the design is about.


So will it help Randy find his culprit,

We can only hope,

This time, he’ll find Anton all right,

And lead to the end of a rope.


Cowboy throwing lasso

If Randy follows Anton over time,

And cannot afford to miss rhythm and rhyme,

Then what with missing values to do?

Don’t laugh for it could be you.


Perhaps a pattern mixture model is in order,

And will not make you run for the border,

To predict where Anton is on his way,

And at least have Randy save the day?


Save the day? Y’know like Mighty Mouse?  No? Okay, I tried though.



Because why not??


So if two worlds are different,

How do you know?

What is the sizes are too small?

For the difference to show?


So how big of a world does Randy need.

To find it different between the good and bad.

And might it finally lead to Anton?

Or is it another confusion to be had?



Simple yet interesting topic that was covered at the Joint Statistical Meetings this week.


So what is the best way to choose a plot?

If regression, do we chose a line?

Or survival, a curve?

Would that be just fine?


Perhaps the scientists working with Randy could these use.

When and how to best catch Anton to him to show.

But such a choice may not be as simple as it seems.

And now you are in the know.


And since the Joint Statistical Meetings were in Chicago, here is a picture of my hometown 🙂