Q: How many rainforests are there in the world?
A: A census found there’s a brazillion.
Q: How many rainforests are there in the world?
A: A census found there’s a brazillion.
Lyrics © 2018 Lawrence M. Lesser
may sing to the tune of "Annie's Song" by John Denver
Don’t try for a census
when the best thing’s a sample.
But now how could a sample
surpass trying for all?
Well, samples are faster
(so less outdated data)
and samples are cheaper
(less workers to hire)!
Sometimes a census
simply can’t even be done:
If you needed a blood test,
should they take some or all?
If they crash-tested all cars,
they’d have no cars to sell you!
Just take a sample
from what is mixed well.
If the sample’s much smaller
than the whole population,
Then precision goals lean on
the sample size n,
Not the sampling fraction:
that is really amazing!
Don’t try for a census,
come sample instead!
Q: What do you call noting prices of the candy by the front door at your neighborhood 7-Eleven?
A: Convenience sampling.
Larry Lesser
What explains why some Senators ignore testimony when they vote on Federal Court nominations?
Answer: Confirmation bias.
Larry Lesser and Dennis Pearl
lyric © 2017 Lawrence M. Lesser
sing to the tune of the KC & the Sunshine Band hit “(Shake, Shake, Shake) Shake Your Booty”
Two…. million readers
picked Landon to
advance;
Gallup
disagreed
from thousands picked by chance.
CHORUS:
(Shake, shake, shake,
shake, shake, shake:
random sample,
random sample)2x
Ahhh, you think
you can pick them
very well,
But drafting
for Vietnam
wasn’t swell.
(Repeat Chorus)
Shake, shake, shake
shake, shake, shake:
iPod Shuffle,
quick pick lotto.
Shake, shake, shake,
take, take, take:
random sample!
Lyrics © 2016 Amy Adler
May be sung to the tune of "Miss Susie Had a Steamboat"
Oh, once there was a teacher who wondered 'bout his class.
(Do they) smile at him to please him? Or like to learn 'bout stats?
So he set out to determine but didn't want to ask
Every single student -- that was too much of a task.
Instead he aimed to sample and interview some instead.
But how to do that properly? He thought out in his head.
He had to choose some students to represent the whole:
What things should he consider? What does he need to know?
How should he word the questions? And are there other facts,
That might affect the outcome? He needs to know all that.
Once he knows the questions and other relevant facts,
He can take a sample that represents the class.
But picking only certain ones, the result may go astray
Picking those who get straight A's may be a biased way.
Those who get straight A's might be more
Prone to answer in the same way and skew the results for sure.
To truly have the answer with high confidence,
A random sample when you can is generally the best.
So once he picked the sample -- he'd ask them one by one,
And then he'd calculate how many said "It's fun!"
He didn't want to call them or ask them face to face,
So he emailed and repeated 'till every poll took place.
He looked at the percentage who said that stats are great,
And felt the un-polled students should agree at that same rate.
So if he took a sample that represents the whole:
He can estimate how many are stats fans on his roll.