Super Sample Size Me
From Y to X-cellent
Count your Tukey-Duckworth's
Words Matter
100

Increased sample size, _____ Confidence Intervals

What is Decreased?

100

Functional, Structural, Temporal - Oh My!

What are Y to X split types?

100

The mathematical basis of Tukey's methodology, making it robust

What is Probability?

100

"It’s amazing the things you can see when you ____" - Yogi Berra

What is "look"?

200

Know Your Data - The formula you use to determine sample size depends on this basic determination of your data/system

What is your data type?

Continuous vs Categorical (Binomial, Poisson, Rare Event)

200

Y -> X : Progressive search for ______

What is Cause?

Not Solution

200

You've plotted your HI & LOW data, now count this

What is Endcount?

200

Number of repeats within the subgroup (k)

What is n?

300

Don't want to be caught behind the 8 ball, you know when you should get your sample size

What is during experimental design?

300

Wait! Did you skip the first step?

You have to confirm this, and put it on your Y to X to show you completed it

What is an MSA?

Repeatability & Reproducibility

300

Endcount of the "Secret Handshake"

What is 6?

3 x HI & 3 X LO, no overlap

300

The +/- margin of error around your sample estimate

What is confidence interval?

400

Non-Homogeneity happens, therefore 2 types of sample sizes should be considered

What is n & k?

n - # of repeats within each subgroup (k)

k - # of replicates of the subgroups

400

Try to identify this pair as soon as possible - causal category with distinctly different results in The Y

What is a Diagnostic Pair?

400

A narrower distribution of the Y (more overlap) requires this

What is a larger sample size?

400

Assignable vs Common

Cause that creates a ‘sudden’ change in a stable process; either in a nonrandom pattern or beyond predictable (historical) limits.

What is Assignable?

500

The minimum number of subgroups per condition to support that a difference exists with 95% confidence

What is 3?

Tukey-Duckworth Endcount Method - The probability of the 3 good lots being lower than the 3 bad lots by chance alone is 5%.

500

The Effect to Cause search can start with one of these "Four Questions"

What is:

What's Wrong?

What's Changed?

What's Different?

What's Happening?

500

The probability of a "False Positive"

What is alpha?

500

An output characteristic that varies and can be Measured or Assessed (can be in an acceptable or unacceptable state)

What is The Y?

M
e
n
u