2_sample_power - The power of the two-sample mean...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
ET , where ET = μ x - μ y q σ 2 x /n x + σ 2 y /n y = q n x + n y μ x - μ y q σ 2 x /p x + σ 2 y /p y = q n x + n y μ x - μ y q σ 2 x /p x + σ 2 y / (1 - p x ) Suppose the raw effect size μ x - μ y , the total sample size n x + n y , and the variances σ 2 x and σ 2 y are all fixed, and we want to maximize the power by setting the value of p x to its ideal point. This becomes equivalent to minimizing σ 2 x /p x + σ 2 y / (1 - p x ) . Note that the minimum of this expression cannot occur on the boundary ( p x = 0 or p x = 1), since it becomes equal to + in either case. If we differentiate with respect to p x , we get - σ 2 x /p 2 x + σ 2 y / (1 - p x ) 2 , so the minimum occurs when - σ 2 x /p 2 x + σ 2 y / (1 - p x ) 2 = 0 . Thus we have - σ 2 x (1 - p x ) 2 + σ 2 y p 2 x = 0 , which simplifies to ( σ 2 y - σ 2 x ) p 2 x + 2 σ 2 x p x - σ 2 x = 0 . This is quadratic in
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 02/06/2012 for the course STAT 403 taught by Professor Kerbyshedden during the Winter '12 term at University of Michigan-Dearborn.

Page1 / 2

2_sample_power - The power of the two-sample mean...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online