Unformatted text preview: SECTION 7.7. The Operating Characteristic Curves Bring Table 8 (P.585 and p.586) and Table 3(P574575). 1. OC CURVE : Graph of L( μ )= P(Accept Ho for mean = μ  sample size=n and α ). . L( μ ) = β ( μ ) if μ is a value for which H 1 is true. = 1 α if μ = μ > 1 α if μ is a value for which H 0 is true. Example: Compute the value of the OC curve for n=4 α =0.05 two sided for μ =6 for the case where μ =12, 6, 0 and σ =4. Using the result P(Z> c) = P(Z< c) = 1 Φ (c) = Φ (c) . (1) For μ 1 < μ β = P(Z> [ √ n( μ μ 1 )/ σ ] z α }=P(Z<[ √ n( μ 1 μ )/ σ ] +z α })= Φ [Z α( √ n( μ 1 μ )/ σ )]}. (2)For μ 1 > μ 0 β = P[ Z < Z α + ( √ n( μ μ 1 )/ σ ) ] = Φ [ Z α √ n( μ 1 μ )/ σ ] L( μ ) is a function of d=  μ 1 μ / σ = standardized departure from Ho or “effect size ”, n and α wherever μ is a value in H 1 . Note that as d increases L( μ ) decreases for a given n and α ....
View
Full
Document
This note was uploaded on 01/31/2008 for the course AMS 310.01 taught by Professor Mendell during the Fall '03 term at SUNY Stony Brook.
 Fall '03
 Mendell

Click to edit the document details