Conditional Tests for Censored Match Pairs - 1985 Popovich and Rao

17
PLEASE SCROLL DOWN FOR ARTICLE This article was downloaded by: [Popovich, Edward] On: 24 November 2009 Access details: Access Details: [subscription number 917111080] Publisher Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37- 41 Mortimer Street, London W1T 3JH, UK Communications in Statistics - Theory and Methods Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713597238 Conditional Tests for censored matched pairs E.A. Popovich a ; P.V. Rao b a Harris Corporation, Melbourne, FL b Unversity of Florida, Gainesville, FL To cite this Article Popovich, E.A. and Rao, P.V.'Conditional Tests for censored matched pairs', Communications in Statistics - Theory and Methods, 14: 9, 2041 — 2056 To link to this Article: DOI: 10.1080/03610928508829029 URL: http://dx.doi.org/10.1080/03610928508829029 Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

Transcript of Conditional Tests for Censored Match Pairs - 1985 Popovich and Rao

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Popovich, Edward]On: 24 November 2009Access details: Access Details: [subscription number 917111080]Publisher Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Communications in Statistics - Theory and MethodsPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713597238

Conditional Tests for censored matched pairsE.A. Popovich a; P.V. Rao b

a Harris Corporation, Melbourne, FL b Unversity of Florida, Gainesville, FL

To cite this Article Popovich, E.A. and Rao, P.V.'Conditional Tests for censored matched pairs', Communications inStatistics - Theory and Methods, 14: 9, 2041 — 2056To link to this Article: DOI: 10.1080/03610928508829029URL: http://dx.doi.org/10.1080/03610928508829029

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

COMMUN. STATIST.-THEOR. METH., 1 4 ( 9 ) , 2 0 4 1 - 2 0 5 6 (1985)

E. A. Popov ich

t k r r i s C o r p o r a t i o n ' Melbourne, FL 32919

CONDITIONAL TESTS FOR CENSORED MATCHED PAIRS

a n d P. V . Rao

U n i v e r s i t y of F l o r i d a G a i n e s v i l l e , FL 32611

ABSTRACT

A c l a s s o f s t a t i s t i c s i s p roposed f o r t h e problem of t e s t i n g

f o r l o c a t i o n d i f f e r e n c e u s i n g randomly r i g h t c e n s o r e d matched p a i r

d a t a . Each member of t h e c l a s s p r o v i d e s a c o n d i t i o n a l l y d i s t r i b u -

t i o n - f r e e t e s t o f Ho: n o l o c a t i o n d i f f e r e n c e . S i m u l a t i o n r e s u l t s

i n d i c a t e t h a t powers o f t e s t s based o n c e r t a i n members i n t h e

c l a s s a r e a s good a s o r b e t t e r t h a n t h e power o f a t e s t proposed

by Woolson a n d Lachenbruch (1980).

INTRODUCTION

The problem of t e s t i n g f o r d i f f e r e n c e between two t r e a t m e n t s

u s i n g randomly r i g h t c e n s o r e d matched p a i r d a t a was examined by

Woolson a n d Lachenbruch (1980) among o t h e r s . Under t h e a s sumpt ion

t h a t b o t h members of a p a i r have e q u a l b u t random o b s e r v a t i o n

t i m e s , Woolson and Lachenbruch deve loped a f a m i l y of t e s t

s t a t i s t i c s f o r t e s t i n g t h e n u l l h y p o t h e s i s Ho: No d i f f e r e n c e i n

Copyright @ 1985 by Marcel Dekker, Inc.

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

2042 POPOVICH AND RAO

t r e a t m e n t s . The Woolson-Lachenbruch (WL) s t a t i s t i c s u t i l i z e t h e

concep t of g e n e r a l i z e d r a n k v e c t o r i n t r o d u c e d by K a l b f l e i s c h a n d

P r e n t i c e (1973) . These s t a t i s t i c s were d e r i v e d by f o l l o w i n g t h e

d e r i v a t i o n of t h e l o c a l l y most power fu l (LMP) r a n k t e s t f o r

l o c a t i o n i n t h e uncensored matched p a i r s c a s e . Woolson and

I achenbruch p r e s e n t two v e r s i o n s of t h e i r s t a t i s t i c : one i s b a s e d

o n d o u b l e e x p o n e n t i a l s c o r e s a n d t h e o t h e r b a s e d o n l o g i s t i c

s c o r e s .

I n s p i t e of t h e f a c t t h a t t h e WL-tes ts p r o v i d e a n i n t u i t i v e l y

r e a s o n a b l e method f o r t e s t i n g Ifo, t h e i r c o m p l i c a t e d f u n c t i o n a l

form makes i t d i f f i c u l t t o i n v e s t i g a t e t h e i r t h e o r e t i c a l

p r o p e r t i e s . For example , i n v iew o f t h e n o t e by Woolson and

Lachenbruch ( 1 9 8 4 ) , i t i s n o t c l e a r t h a t t h e s e t e s t s a r e LMP i n

t h e c e n s o r e d c a s e . Tne proof of t h e i r a s y m p t o t i c n o r m a l i t y , found

i n t h e WL-paper, i s s t r i c t l y v a l i d o n l y when N l , t h e number of

uncensored p a i r s , a n d N 2 , t h e number of s i n g l y c e n s o r e d p a i r s , a r e

r e g a r d e d a s nonrandom w i t h N , + N 2 t e n d i n g t o i n f i n i t y . It i s n o t

known whe the r t h e a s y m p t o t i c n o r m a l i t y h o l d s u n c o n d i t i o n a l l y a s n ,

t h e t o t a l number o f p a i r s , t e n d s t o i n f i n i t y .

For s m a l l sample s i z e s e x a c t c r i t i c a l v a l u e s f o r t h e WL-test

c o u l d b e b a s e d on t h e i r c o n d i t i o n a l p e r m u t a t i o n d i s t r i b u t i o n ,

c o n d i t i o n a l on t h e o b s e r v e d c e n s o r i n g p a t t e r n . However, t h e

d e t e r m i n a t i o n o f t h e s e c r i t i c a l v a l u e s becomes p r o g r e s s i v e l y

i m p r a c t i c a l when max(N1 ,N2) > 8 and min(N1,N2) > 2 .

The o b j e c t i v e i n t h i s p a p e r i s t o p ropose a n a l t e r n a t i v e

c l a s s of s t a t i s t i c s , r e f e r r e d t o a s Cn i n t h e s e q u e l , f o r t e s t i n g

h0. As w i l l b e s e e n f r o m what f o l l o w s , t h e s t a t i s t i c s i n cn w i l l

b e c o m p u t a t i o n a l l y s i m p l e and w i l l n o t s u f f e r f rom t h e drawbacks

of t h e W L - s t a t i s t i c s . F o r example, a Tn E ( f n w i l l be

a s y m p t o t i c a l l y normal a s n + m under s u i t a b l e r e g u l a r i t y

c o n d i t i o n s , and c o n d i t i o n a l on t h e o b s e r v e d v a l u e s of N 1 a n d N2

w i l l b e d i s t r i b u t i o n - f r e e . Because t h e s e c o n d i t i o n a l n u l l

d i s t r i b u t i o n s a r e d i s t r i b u t i o n s o f l i n e a r c o m b i n a t i o n s o f

i n d e p e n d e n t b i n o m i a l and Wilcoxon s i g n e d r a n k s t a t i s t i c s , t h e

c r i t i c a l v a l u e s of t h e s t a t i s t i c s i n (f a r e e a s i l y t a b l u l a t e d f o r

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

CENSORED MATCHED P A I R S 2043

a l l r e a s o n a b l e v a l u e s o f N1 a n d N 2 . F u r t h e r m o r e , s i m u l a t i o n

r e s u l t s i n d i c a t e t h a t t h e powers of c e r t a i n members o f cn a r e a s

good a s , and i n some c a s e s b e t t e r t h a n , t h e power of t h e WL-

s t a t i s t i c s .

2. N O T A T I O N a n d PRELIMINARY RESULTS

C o n s i d e r a n expe r imen t i.n which two t r e a t m e n t s a r e a p p l i e d t o

e a c h of t h e n matched p a i r s of e x p e r i m e n t a l u n i t s r e s u l t i n g i n t h e

r e s p o n s e (x? x0 ) , i = 1,2, ..., n. L e t Ci b e t h e v a l u e of a 11' 1 2

c e n s o r i n g v a r i a b l e common f o r t h e i t h p a i r . The o b s e r v e d r e s p o n s e s

f o r t h e i t h p a i r a r e (Xil ,Xi2) and (6i1, di2) where X i j = min(XYj,

Ci) a n d 8 . . = I o r U a c c o r d i n g a s X. = X Y j o r C i . I n what 1~ l j

t h e a s sumpt ions A1 - A 5 l i s t e d b e l o *

i a t e .

f o l l o w s , v a r i o u s s u b s e t s of

w i l l be u t i l i z e d a s appropr

A l : (xY1, x:~) a r e i n d e p e r [ d e n t i d e n t i c a l l y d i s t r i b u t e d ( i . i .d.)

a s a c o n t i n u o u s b i v a r i a t e random v a r i a b l e (x:, x:).

A2: There e x i s t s a r e a l number 0 s u c h t h a t (x:, X; + 0) i s a n

exchangeab le p a i r .

A3: Ci i = 1 , 2 , . ..,n a r e i . i . d . a s a c o n t i n u o u s random v a r i a b l e

C. A l so , C a n d (x? , x;) a r e r m t u a l l y i n d e p e n d e n t .

A 4 : Ci a n d (XP1, XY2) i = 1 , 2 , . . . ,n a r e m u t u a l l y independen t .

A 5 : Let P I . } d e n o t e p r o b a b i l i t y c a l c u l a t e d w i t h t h e v a l u e o f 0 0

s a t i s f y i n g A2. Then P ~ { x : > C , i = 1 , 2 ) < 1.

L ike t h e W L - s t a t i s t i c s , t h e s t a t i s t i c s we wish t o c o n s i d e r

a r e b a s e d o n t h e o b s e r v e d d i E f e r e n c e s Di = X i l - Xi2. A

d i f f e r e n c e Di w i l l b e s a i d t o b e l o n g t o c a t e g o r y C1 i f bo th Xil

a n d Xi2 a r e uncensored , t o c a t e g o r y C2 i f e x a c t l y one o f Xil a n d

Xi2 i s uncensored , and t o c a t e g o r y C3 i f b o t h Xil and Xi2 a r e

c e n s o r e d . As n o t e d by Woolson a n d J a c h e n b r u c h (1980) a Di i n C1

w i l l e q u a l t h e t r u e d i f f e r e n c e D: = xY1 - x:*, whereas a Di i n C2

w i l l imply t h a t t h e t r u e d i f f e r e n c e DP i s l e f t o r r i g h t c e n s o r e d

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

2044 POPOVICH AND RAO

a t t h e o b s e r v e d d i f f e r e n c e Di a c c o r d i n g a s 6il = 1 o r 0 . The

o r d e r r e l a t i o n s h i p be tween Di a n d D; f o r t h o s e Di i n C3 i s

unknown.

Suppose f o r k = 1 , 2 , and 3 , Nk d e n o t e s t h e o b s e r v e d number of

p a i r s i n c a t e g o r y Ck, a n d l e t (Xilk, XiZk), xi = Xilk - Xi2k, i =

1 , 2 , . . . ,Nk d e n o t e t h e s e p a i r s a n d t h e i r d i f f e r e n c e s r e s p e c t i v e l y .

C l e a r l y we h a v e ,

B e f o r e p r o c e e d i n g f u r t h e r , some comments c o n c e r n i n g

a s s u m p t i o n s A1-A5 a r e i n o r d e r . From a s s u m p t i o n A2 i t f o l l o w s

t h a t XO a n d XO + 0 h a v e i d e n t i c a l m a r g i n a l d i s t r i b u t i o n s . 1 2

Assumption A3 r e q u i r e s t h a t b o t h X: and XO have a common 1 1 1 2

c e n s o r i n g t i m e Ci which may be a n u n r e a l i s t i c a s s u m p t i o n i n some

a p p l i c a t i o n s . I f C . i s t h e c e n s o r i n g t i m e f o r xo t h e n o u r 1 j i j '

model would b e a p p r o p r i a t e i f Ci = min (Ci1,Ci2) i s r e g a r d e d a s 0 0

t h e common c e n s o r i n g t i m e f o r ( x ~ , , X ~ ~ ) . Assumpt ion A5 i m p l i e s

PO{N1 + N2 > 0) > 0 , s o t h a t w i t h p o s i t i v e p r o b a b i l i t y t h e r e w i l l

be a t l e a s t one p a i r i n c a t e g o r y C1 o r C2. Obv ious ly , i f

Pe{Nl + N 2 > 0 ) = 0 , t h e n e v e r y p a i r b e l o n g s t o c a t e g o r y C3 a n d

t e s t s b a s e d on o b s e r v e d d i f f e r e n c e s Di w i l l n o t be i n f o r m a t i v e .

If P0(N2 = N 3 = 0 ) = 1, t h e n t h e problem u n d e r c o n s i d e r a t i o n

i s t h e u s u a l uncensored matched p a i r s problem a n d t h e r e e x i s t a

number of l i n e a r s i g n e d r a n k t e s t s (Theorem I1 4.9 , Hajek ( 1 9 6 7 ) )

f o r t e s t i n g Ho:O = 0. The well-known Wilcoxon s i g n e d r a n k t e s t i s

a n i m p o r t a n t example . However, i f P0{N2 > 0) > 0, t h e n t h e

s i t u a t i o n i s d i f f e r e n t b e c a u s e t h e t r u e d i f f e r e n c e DP between a

p a i r i n C2 w i l l be r i g h t o r l e f t c e n s o r e d a t t h e o b s e r v e d

d i f f e r e n c e D i , and , c o n s e q u e n t l y , n o t a l l o b s e r v e d d i f f e r e n c e s

w i l l e q u a l t h e t r u e d i f f e r e n c e s . We s h a l l u s e t h e symbols N2= and

N2R t o d e n o t e , r e s p e c t i v e l y , t h e number o f l e f t - c e n s o r e d and t h e

number of r i g h t - c e n s o r e d d i f f e r e n c e s . Note t h a t

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

CENSORED MATCHED PAIRS

N2 = N Z L + N 2 R , N Z L = E6i1(l-6i2) a n d N Z R = C(1-6 ) 6 i l 1 2 '

Then Pk(B) d e n o t e s t h e p r o b a b i l i t y t h a t a n o b s e r v e d p a i r f a l l s

i n t o c a t e g o r y Ck, k = 1 ,2 , and , PZL(B) a n d P2R(8) d e n o t e ,

r e s p e c t i v e l y , t h e p r o b a b i l i t y t h a t a n o b s e r v e d p a i r i s l e f t -

c e n s o r e d o r r i g h t - c e n s o r e d .

Lemma 2.1: Suppose a s s u m p t i o n s Al , A2, A3, a n d A4 h o l d .

( a ) If 6 = 0 a n d P l ( 0 ) > 0 , t h e n c o n d i t i o n a l o n N l = n1 > 0 ,

D l l , D12,...,D h a v e i n d e p e n d e n t i d e n t i c a l d i s t r i b u t i o n s In1

symmet r i c a b o u t 0.

( b ) I f P 2 ( 8 ) > 0, t h e n t h e c o n d i t i o n a l d i s t r i b u t i o n of NZR,

c o n d i t i o n a l o n N2 = n2 > 0 , i s b i n o m i a l w i t h p a r a m e t e r s n2

a n d PZK(8) /P2(8 ) .

( c ) C o n d i t i o n a l o n Ni = ni , i = l , 2 , 3 , t h e random v a r i a b l e s

DlI,D12,.. . ,D a n d NZR a r e r m t u a l l y i n d e p e n d e n t . In1

P roof o f Lemma ( 2 . 1 ) , wh ich i s s t r a i g h t f o r n a r d , w i l l b e l e f t o u t

f o r b r e v i t y . Lemma ( 2 . l a ) s u g g e s t s t h a t t h e obse rved d i f f e r e n c e s

i n Cl c o u l d be u t i l i z e d t o d e f i n e a c o n d i t i o n a l t e s t o f Ho: 8=O.

C l e a r l y , a t e s t i n t h e f a m i l y of l i n e a r s i g n e d r a n k t e s t s (Ha jek ,

1967) c a n b e s e l e c t e d o n t h e b a s i s of a n assumed d e n s i t y f o r

DIj . However, i n t h i s p a p e r we s h a l l r e s t r i c t o u r a t t e n t i o n t o

t h e Wilcoxon s i g n e d r a n k t e s t b e c a u s e o f i t s h i g h e f f i c i e n c y f o r a

wide v a r i e t y of d i s t r i b u t i o n s a n d t h e a v a i l a b i l i t y of t a b l e s o f

c r i t i c a l v a l u e s f o r a l a r g e s e l e c t i o n of sample s i z e s .

Accord ing ly we c o n s i d e r t h e s t a t i s t i c

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

2046 POPOVICH AND RAO

where , + ( u ) = 1 o r 0 a c c o r d i n g a s u > 0 o r < 0 and R; i s t h e r a n k

of 1 ~ 1 ~ 1 among I D I I I ~ I D ~ ~ I ~ ~ ~ ~ I D ~ N ~ I *

From Lemma (2. l b ) , i t f o l l o w s t h a t u n d e r Ho: 0=0, t h e

c o n d i t i o n a l d i s t r i b u t i o n o f N 2 R i s b i n o m i a l w i t h p a r a m e t e r s n2 and

L/2. T h e r e f o r e , a s e c o n d t e s t o f Ho:8 = 0, u t i l i z i n g d i f f e r e n c e s

i n C2, c o u l d b e b a s e d o n t h e s t a t i s t i c

C l e a r l y , u n d e r Ho:B=O, t h e c o n d i t i o n a l t e s t s based o n Tln a n d TZn

a r e d i s t r i b u t i o n - f r e e a n d f r o m Lemma (2 . l c ) t h e s e t e s t s a r e

i n d e p e n d e n t . h u s , a n y method o f combining two i n d e p e n d e n t t e s t s

c o u l d b e u t i l i z e d t o d e v e l o p a d i s t r i b u t i o n - f r e e c o n d i t i o n a l t e s t

o f Ho:8=0. C r i t i c a l v a l u e s f o r t h e s e t e s t s a r e e a s i l y t a b u l a t e d

by u t i l i z i n g t h e c r i t i c a l v a l u e s f o r i n d e p e n d e n t Wilcoxon-signed

r a n k a n d b i n o m i a l t e s t s . I n S e c t i o n 3 we c o n s i d e r a c l a s s o f

s t a t i s t i c s formed by c e r t a i n l i n e a r combina t ions o f Tln and T2n.

Throughout h e r e a f t e r , c o n d i t i o n a l d i s t r i b u t i o n s r e f e r t o

d i s t r i b u t i o n s c o n d i t i o n a l o n Ni = n i , i = l , 2 , 3 . It i s c l e a r t h a t

u n d e r Ho:8=0, t h e c o n d i t i o n a l mean and v a r i a n c e of Tin a r e ,

r e s p e c t i v e l y , pin(ni) a n d u? ( n . ) , i = 1 , 2 where v l n ( n l ) = i n 1

( ) n l ( n l + l ) , ~ ~ ~ ( n ~ ) = O , o f n ( n l ) = ( l / 2 4 ) n l ( n l + l ) ( 2 n l + l ) a n d

u;,(n2) = n2. L e t T;,(N~) d e n o t e t h e s t a n d a r d i z e d v e r s i o n s (unde r

Ho) o f Tin. Tha t i s ,

For t e s t i n g Ho:8 = 0 , we p ropose u s i n g

f o r m

a l i n e a r combina t ion i n t h e

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

CENSORED MATCHED PAIRS 2047

where {L,} i s a sequence o f random v a r i a b l e s s a t i s f y i n g t h e

c o n d i t i o n s

( i ) : 0 4 Ln < 1 ,

( i i ) : Ln = Ln(N,, N2) i s a f u n c t i o n o f N 1 a n d N 2 o n l y ,

a n d

( i i i ) : Ln = L + o ( 1 ) a s n + - f o r some c o n s t a n t L. P

We s h a l l d e n o t e by cn t h e c l a s s of s t a t i s t i c s d e f i n e d by (3 .2) .

It i s e a s i l y s e e n t h a t u n d e r Ho, e v e r y member of cn h a s

c o n d i t i o n a l mean a n d s t a n d a r d d e v i a t i o n e q u a l t o 0 a n d 1

r e s p e c t i v e l y , a n d t h a t i t s c o n d i t i o n a l d i s t r i b u t i o n i s symmet r i c

a b o u t 0. S i n c e t h e c o n d i t i o n a l n u l l d i s t r i b u t i o n of Tn(N1,N2) i s

a c o n v o l u t i o n o f independen t Wilcoxon s i g n e d r a n k and b i n o m i a l

d i s t r i b u t i o n s , t h e c r i t i c a l v a l u e s f o r c o n d i t i o n a l t e s t s b a s e d o n

t h e s t a t i s t i c s i n On a r e e a s i l y o b t a i n e d f o r s m a l l t o modera t e

v a l u e s of n i , i = 1 ,2 . 'Lheorem 3.1 s t a t e s c o n d i t i o n s unde r which

Tn E On will b e a s y m p t o t i c a l l y normal.

'Iheorem 3.1. If a s s u m p t i o n s A1 - A 5 h o l d , t h e n f o r e v e r y Tn in

On and r e a l number x,

where a ( * ) i s t h e cd f of a s t a n d a r d normal random v a r i a b l e .

A p roof o f Theorem 3 .1 may b e based on a n a d a p t a t i o n o f

'Iheorem 1 of Anscombe (1952) s t a t e d a s Lemma 3.1. Proof o f Lemma

3.1 , wh ich i s s t r a i g h t f o r w a r d , may b e found i n Popovich ( 1 9 8 3 ) .

Lemma 3.1. L e t {Tn 1 , n l = 1 , 2 , . . . , n2= 1 ,2 , . . . , b e a n a r r a y o f 1 '"2

random v a r i a b l e s s a t i s f y i n g c o n d i t i o n s ( i ) a n d ( i i ) .

CONDITION ( i ) : There e x i s t s a cd f F ( * ) s u c h t h a t a t e v e r y

c o n t i n u i t y p o i n t x of F ( * ) Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

2048 POPOVICH AND RAO

CONDITION ( i i ) : C o r r e s p o n d i n g t o e a c h E , T- > 0, t h e r e e x i s t

c o n s t a n t s v a n d c s u c h t h a t m i n ( n l , n 2 ) > v i m p l i e s

( n f - n i ) < c n i , i = 1 ' 2 ) > 1- q .

} i s a s e q u e n c e o f i n t e g e r v a l u e d random v a r i a b l e s s u c h n t h a t p-l im n-lNin = A. A . > O , i = 1 , 2 , t h e n

1' 1

a t e v e r y c o n t i n u i t y p o i n t x o f F ( - ) .

A s k e t c h o f a p r o o f o f Theorem 3 . 1 . L e t ~ : , ( n ~ ) r e f e r t o t h e

random v a r i a b l e T ~ , ( N ~ ) c o n d i t i o n a l o n Ni = n i , i = 1 , 2 . C l e a r l y

TZn(ni) i s a s y m p t o t i c a l l y n o r m a l , a s n . + -, a n d by Lemma 2 . l ( c ) ,

~ : , ( n ~ ) a n d T i n ( n 2 ) a r e m t u a l l y i n d e p e n d e n t . T h e r e f o r e ,

i s a s y m p t o t i c a l l y n o r m a l a s m i n ( n l , n 2 ) + m. T h a t i s , {T;1,n2}

s a t i s f i e s CONIIITION ( i ) o f Lemma 3.1. Now s i n c e f o r ni + m, we

h a v e t h e r e p r e s e n t a t i o n ~ q , ( n ~ ) = ~ Z ~ ( n . 1 + o ( I ) , w h e r e LlZn(ni) P

i s a s t a n d a r d i z e d U - s t a t i s t i c b a s e d o n ni o b s e r v a t i o n s , we c a n

u t i l i z e t h e r e s u l t o f S p r o u l e ( 1 9 7 4 ) t o show t h a t ~ q ~ ( n ~ )

s a t i s f i e s c o n d i t i o n C2 o f 4nscombe ( 1 9 5 2 ) f o r i = 1 , 2 . The r m t u a l

i n d e p e n d e n c e o f ~ : ~ ( n ~ ) a n d ~ ; ~ ( n ~ ) i m p l i e s , a f t e r some a l g e b r a ,

t h a t {TG 1 s a t i s f i e s CONIIITION ( i i ) o f Lemma 3.1. Note t h a t l '"2

u n d e r Ho, P i ( 0 ) = p - l i m n - ' ~ ~ a s n + m f o r i = 1 , 2 a n d by A 5 Pi(U)

> 0 f o r a t l e a s t o n e i , i = 1 , 2 . T h e r e f o r e , f r o m Lemma 3.1, i t

f o l l o w s t h a t u n d e r HO, T i l r N 2 i s a s y m p t o t i c a l l y n o r m a l a s n + m.

Be p r o o f t h a t TN i s a s y m p t o t i c a l l y n o r m a l u n d e r HO 1 ' 2

f o l l o w s b e c a u s e a s n + -

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

CENSORED MATCHED PAIRS 2049

Theorem 3.1 shows t h a t f o r l a r g e v a l u e s of n, a n uncondi-

t i o n a l t e s t b a s e d on t h e a s y m p t o t i c n o r m a l i t y o f Tn(N1,N2) c a n b e

u s e d f o r t e s t i n g HO. I n S e c t i o n 4 we c o n s i d e r t h e problem o f

s e l e c t i o n o f a s t a t i s t i c f rom En.

4 . A STATISTIC FOR TESTING Ho

S i n c e e v e r y s t a t i s t i c i n En is a l i n e a r combina t ion of two

s t a t i s t i c s which a r e c o n d i t i o n a l l y i n d e p e n d e n t a n d a s y m p t o t i c a l l y

normal , i t i s n a t u r a l t o c o n s i d e r t h e p o s s i b i l i t y of s e l e c t i n g a

s t a t i s t i c f rom K n b a s e d on c o n d i t i o n a l P i tman A s y m p t o t i c R e l a t i v e

E f f i c i e n c y (PARE) a s min (n l ,n2) + =. Theorem 10.2.3 of S e r f l i n g

(1980) c a n be used t o d e r i v e a n e x p r e s s i o n f o r L s u c h t h a t t h e

s t a t i s t i c TnL = ( l - ~ ) ' ~ T* ( n ) + Lm5 Tk ( n ) has 'Ondi- In 1 2n 2

t i o n a l PARE i n On. U n f o r t u n a t e l y t h e form of Tnl, depends on t h e

j o i n t d i s t r i b u t i o n o f (x~,x: ,c) i n a c o m p l i c a t e d manner a n d o u r

a t t e m p t s t o s i m p l i f y t h e e x p r e s s i o n a r e n o t s u c c e s s f ~ l l t h u s f a r .

A l t e r n a t e l y , o n e c a n s e l e c t s p e c i f i c l i n e a r combina t ions i n

and examine t h e i r PARE'S r e l a t i v e t o t h e W L - s t a t i s t i c s . I n

s p i t e of t h e f a c t t h a t e f f i c a c i e s o f s t a t i s t i c s i n Q n c a n he

d e r i v e d i n a s t r a i g h t - f o r w a r d manner t h e d e r i v a t i o n o f PARE'S w i t h

r e s p e c t t o t h e W L - s t a t i s t i c s d o e s n o t a p p e a r t o be a n e a s y t a s k .

The d i f f i c u l t y h e r e i s due t o t h e t h e o r e t i c a l l y c o m p l i c a t e d

f u n c t i o n a l forms o f W L - s t a t i s t i c s . Consequen t ly , we r e s o r t e d t o a

s i m u l a t i o n s t u d y t o compare f o u r s p e c i f i c s t a t i s t i c s i n cn w i t h

T-WL, t h e X L - s t a t i s t i c b a s e d on l o g i s t i c s c o r e s . The W L - s t a t i s t i c

based on d o u b l e e x p o n e n t i a l s c o r e s was n o t i n c l u d e d i n o u r s t u d y

mainly b e c a u s e i n t h e uncensored c a s e , t h e s i g n e d r a n k t e s t i s

g e n e r a l l y p r e f e r r e d o v e r t h e s i g n t e s t f o r a v a r i e t y o f e r r o r

d i s t r i b u t i o n s .

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

2050 POPOVICH AND RAO

When s e l e c t i n g s t a t i s t i c s from %, primary c o n s i d e r a t i o n was

g iven t o t h e u s u a l methods of forming l i n e a r combinat ions of two

s t a t i s t i c s . The f i r s t s t a t i s t i c s e l e c t e d from c, corresponds t o

e q u a l weigh ts , i .e. Ln = .5:

I f Ln i s t a k e n p r o p o r t i o n a l t o N 2 , t h a t i s , Ln= N ~ ( N ~ + N ~ ) - . ~ ,

we g e t t h e s t a t i s t i c

T-SQK = (N1+ N ~ ) - . ~ [ N . ~ ~ T ; ~ ( N ~ ) + N .52 T ; ~ ( N ~ ) ] ,

w h i l e Ln p r o p o r t i o n a l t o N; y i e l d s t h e s t a t i s t i c

T-SS = (N: + N $ ) - ' ~ [ N ~ T ~ , ( N ~ ) + N ~ T ; ~ ( N ~ ) I .

F i n a l l y , s e l e c t i n g Ln p r o p o r t i o n a l t o o2 (N ) , t h a t i s wi th 2n 1

- 1 Ln = a ; n ( ~ l ) [o tn(Nl ) + O ; ~ ( N ~ ) ] , we g e t t h e f o u r t h s t a t i s t i c

Fol lowing Woolson and Lachenbruch (l98O), we assumed t h e l o g

l i n e a r model:

where I$ > O i s a n unknown parameter , V i l and V i 2 a r e i . i . d .

nonnegat ive random v a r i a b l e s , and W . i s a n independent nonnegat ive

random v a r i a b l e f o r a l l i. It i s c l e a r t h a t i f 8 = l o g I$ t h e n ,

0 0 Xil - Xi2 = 9 + ( l o g Vil - l o g Vi2), i = 1 ,2 , .. . ,n;

0 0 consequent ly , t h e d i s t r i b u t i o n a l form of X i l - Xi2 depends on t h e

d i s t r i b u t i o n a l form of l o g Vil - l o g Vi2.

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

CENSORED MATCHED PAIRS 2051

0 0 Four d i s t r i b u t i o n a l forms f o r XiL - Xi2 were s i m u l a t e d i n

t h i s s t u d y . n e l o g i s t i c d i s t r i b u t i o n was s i m u l a t e d by g e n e r a t i n g

V i l a n d V i 2 w i t h i n d e p e n d e n t e x p o n e n t i a l d i s t r i b u t i o n s . Ano the r

l i g h t - t a i l e d d i s t r i b u t i o n , t h e no rma l d i s t r i b u t i o n , was s i m u l a t e d

by g e n e r a t i n g l o g V i l a n d l o g V i 2 a s i n d e p e n d e n t s t a n d a r d normal

v a r i a b l e s . Two h e a v y - t a i l e d d i s t r i b u t i o n s , t h e d o u b l e e x p o n e n t i a l

d i s t r i b u t i o n a n d t h e RambergSchmeiser-Tukey (RST) Lambda D i s -

t r i b u t i o n (Rand les and Wolfe, 1979, p. 4 1 6 ) were a l s o s t u d i e d .

The d o u b l e e x p o n e n t i a l d i s t r i b u t i o n was a r r i v e d a t by g e n e r a t i n g

l o g V i l a n d l o g Vi2 w i t h i n d e p e n d e n t e x p o n e n t i a l d i s t r i b u t i o n s .

The KST Lambda c u m u l a t i v e d i s t r i b u t i o n f u n c t i o n (c.d.f .) c a n n o t

be e x p r e s s e d e x p l i c i t l y b u t i t s i n v e r s e c .d . f . i s g i v e n a s

f o l l o w s :

As shown i n Ramberg a n d Schmei se r (1972) when X1 = 0 , Xg = X4 = -1

a n d X 2 = -3.0674, t h e RST Lambda D i s t r i b u t i o n a p p r o x i m a t e s t h e

Cauchy d i s t r i b u t i o n . % i s p a r t i c u l a r c h o i c e of RST Lambda

D i s t r i b u t i o n was u s e d t o g e n e r a t e l o g V i l a n d l o g Vi2 .

To g e n e r a t e t h e c e n s o r i n g random v a r i a b l e , t h e Uniform [O,B]

d i s t r i b u t i o n was u t i l i z e d f o r t h e l o g i s t i c a n d d o u b l e e x p o n e n t i a l

c a s e s , w h i l e t h e n a t u r a l l o g a r i t h m o f t h e Uniform [O,B]

d i s t r i b u t i o n was u s e d i n t h e normal and RST Lambda c a s e s . The

c h o i c e of B was made i n e a c h c a s e t o e n s u r e t h a t , u n d e r Ho, t h e

p r o p o r t i o n o f uncensored p a i r s was a p p r o x i m a t e l y 75% o f t h e t o t a l

sample s i z e w h i l e a p p r o x i m a t e l y 2 0 % of t h e t o t a l sample s i z e

c o n s i s t e d o f p a i r s i n which e x a c t l y one member of a p a i r was

uncensored . Consequen t ly , a p p r o x i m a t e l y 5 % of t h e p a i r s were n o t

u t i l i z e d s i n c e t h e y were p a i r s i n which b o t h members were

c e n s o r e d .

The powers o f t h e t e s t s of Ho b a s e d on t h e f i v e s t a t i s t i c s

were e s t i m a t e d u t i l i z i n g LOO0 random samples of s i z e n = 50 f rom

t h e d i s t r i b u t i o n of xo - xP2 = 0 + l o g V i l - l o g V i a i 1

c o r r e s p o n d i n g t o s e v e r a l c o m b i n a t i o n s o f a v a l u e of f3 and a

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

2052 P O P O V I C H AND R A O

d i s t r i b u t i o n a l fo rm f o r l o g Vil - l o g Vi2. I n o r d e r t o a s s e s s t h e

s t a t i s t i c a l s i g n i f i c a n c e o f t h e e s t i m a t e d d i f f e r e n c e i n powers ,

t h e s t a n d a r d e r r o r of t h e e s t i m a t e d d i f f e r e n c e was a l s o c a l c u l a t e d

f o r e a c h p a i r o f s t a t i s t i c s . I n a d d i t i o n , t o c o n s e r v e s p a c e we

s h a l l d i s c u s s o n l y t y p i c a l r e s u l t s f r o m t h i s s i m u l a t i o n s t u d y .

More d e t a i l s c a n b e o b t a i n e d f r o m t h e f i r s t a u t h o r .

Recause t h e r e s u l t s f o r l o g i s t i c a n d norma l d i s t r i b u t i o n s

were v e r y s i m i l a r , we s h a l l l e a v e o u t t h e normal d i s t r i b u t i o n f r o m

o u r d i s c u s s i o n . 'Ihe s t a t i s t i c s T-STD a n d T-SS a l s o w i l l be l e f t

o u t b e c a u s e t h e power of t h e s e t e s t s were lower t h a n o t h e r s i n

a l m o s t a l l c a s e s e x c e p t t h a t T-SS performed s l i g h t l y b e t t e r t h a n

T-SQB f o r t h e RST Lambda D i s t r i b u t i o n .

Tab le 4.1 shows a summary o f r e s u l t s f o r t h r e e d i s t r i b u t i o n s ,

l o g i s t i c , d o u b l e e x p o n e n t i a l a n d RST Lambda a n d t h r e e s t a t i s t i c s ,

T-WL, T-EQ a n d T-SQR. In e a c h c a s e t h e h y p t h e s i s t e s t e d was Ho: 0

= 0 a g a i n s t Ha: 0 > 0. Based o n t h e a s y m p t o t i c d i s t r i b u t i o n o f

t h e t e s t s t a t i s t i c s , t h e c r i t i c a l v a l u e was chosen t o b e 1.645 i n

e a c h c a s e i n o r d e r t o have a n a p p r o x i m a t e .05 l e v e l t e s t .

An i n s p e c t i o n o f T a b l e 4.1 shows t h a t b a s e d on t h e i r l a r g e

s a m p l e (n=50) power, none o f t h e t h r e e s t a t i s t i c s emerges a s t h e

b e s t s t a t i s t i c f o r a l l c a s e s c o n s i d e r e d . F o r t h e l o g i s t i c

d i s t r i b u t i o n , T-WL a n d T-EQ have s i m i l a r powers a t a l t e r n a t i v e s

c l o s e t o Ho, b u t T-WL i s b e t t e r a t l a r g e r v a l u e s o f 0. F o r t h e

d o u b l e e x p o n e n t i a l d i s t r i b u t i o n , T-EQ a n d T-S QR behave s i m i l a r l y

w i t h b o t h h a v i n g b e t t e r power t h a n T-WL. F o r t h e RST Lambda

D i s t r i b u t i o n , t h e t e s t based o n T-SQR i s d e f i n i t e l y s u p e r i o r .

T h e r e f o r e , o n t h e b a s i s of t h e i r powers , i t a p p e a r s t h a t t h e

performance of t h e s t a t i s t i c T-SQK g e t s p r o g r e s s i v e l y b e t t e r

r e l a t i v e t o t h e o t h e r s a s one moves f r o m l i g h t e r t a i l e d

d i s t r i b u t i o n s t o h e a v i e r - t a i l e d o n e s .

5. A N EXAMPLE

To i l l u s t r a t e t h e u s e o f t h e s t a t i s t i c s d e f i n e d i n S e c t i o n 3

t h e d a t a s e t c o n s i d e r e d by Woolson a n d Lachenbruch (1980) w i l l b e

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

TABLE 4.1 1000 X DWFERENCL IN POWER . . t

DISTRIBUTION LOGISTIC DEXPO RST-LAbBDA w

e 0.00 0.100 0.250 0.500 0.00 0.100 0.250 0.500 0.00 0.40 0.60 0.80 5 r n

T-WL 48 167 555 9 50 4 5 257 704 993 4 3 200 366 527

(T-WL)-(T-EQ) -4 4 36* 5 0" -1 -17* -17* - 2 6 6 2 12

(T-WL)-(T-SQR) -4 2 4* 6 8* 5 I* - 5 -15* -29* -2 -2 -32* -37" -54%

(T-EQ)-(T-SQR) 0 2 0* 3 2* 1 -4 2 -1 2 0 - 8 -38" -39* -66*

Notes: ( 2 ) 7he t o p row g i v e s 1000 X e s t i m a t e d power o f T-WL

(2 ) The v a l u e of a2 e q u a l s t h e v a r i a n c e o f t h e d i s t r i b u t i o n which e q u a l s

n2 /3 f o r t h e L o g i s t i c d i s t r i b u t i o n a n d 2 f o r t h e Double E x p o n e n t i a l

d i s t r i b u t i o n .

( 3 ) A s t e r i s k n e x t t o a number i n d i c a t e s a d i f f e r e n c e g r e a t e r t h a n o r e q u a l t o t w i c e t h e s t a n d a r d e r r o r .

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

2054 POPOVICH AND RAO

c o n s i d e r e d i n t h i s s e c t i o n . The d a t a s e t i s reproduced below.

TABLE 5.1

DIFFERENCE OF THE LOGARITHMS I N SURVIVAL TIME OF THE SKIN GRAFTS

P a t i e n t 1 2 3 4 5 6

'il- 'i2 0.2436 0.3795 1.3550+ 1.2745 0,3747 0.2578

P a t i e n t 7 8 9 10 I I

Xil- Xi2 -0.2624 -0.1542 0.3819 0.6592 0.4055+

Note: + d e n o t e s a r igh t -censored d i f f e r e n c e .

As can b e s e e n from t h e d a t a , N i = 9, N2L = 0, N2R = 2, and

N2 = 2. It f o l l o w s by some e lementa ry c a l c u l a t i o n s t h a t

Tln = 40, T2n = 2, ~ ; ~ ( 9 ) = 2.073, a n d ~ i ~ ( 2 ) = 1.414.

The t e s t s t a t i s t i c Tn(9,2) c a n b e c a l c u l a t e d i f Ln i s

known. For t h e purpose of i l l u s t r a t i o n we s h a l l t a k e Ln = .5,

which i m p l i e s t h a t t h e t e s t w i l l be based on T-EQ. Now,

Comparing t h i s r e s u l t w i t h t h e c r i t i c a l v a l u e s a p p e a r i n g i n t h e

t a b l e s g iven i n Popovich (1983) i t can be s e e n t h a t

s o t h a t t h e observed v a l u e o f 2.466 i s s i g n i f i c a n t a t t h e .O1

l e v e l . A c t u a l l y t h e e x a c t l e v e l a t t a i n e d i s .00488. As shown i n

Woolson and Lachenbruch (1980) t h e W-L t e s t s t a t i s t i c w i t h

l o g i s t i c s c o r e s y i e l d s t h e v a l u e Z = 2.49, which i n d i c a t e s

s i g n i f i c a n c e a t t h e .O1 l e v e l . Also, t h e e x a c t l e v e l f o r t h e W-L

t e s t s t a t i s t i c based on t h e p e r m t a t i o n d i s t r i b u t i o n o f t h e

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

CENSORED MATCHED PAIRS 2055

l o g i s t i c s c o r e s i s 11 /2O48 o r 0.00537, a g a i n showing s i g n i f i c a n c e

a t t h e .O1 l e v e l . Note t h a t t h e v a l u e of 8/2048 a s g i v e n i n

Woolson and Lachenbruch (1980) i s i n c o r r e c t due t o a minor

computa t iona l e r r o r .

6. C ONCLUS IONS

The s t a t i s t i c s T-EQ a n d T-SQR proposed i n S e c t i o n 4 have

s e v e r a l advantages when a p p l i e d t o randomly r i g h t censored p a i r e d

da ta . They prov ide s imple d i s t r i b u t i o n - f r e e t e s t s based on e a s i l y

t a b u l a t e d c r i t i c a l v a l u e s ( a s opposed t o p e r n u t a t i o n t e s t s w i t h

c r i t i c a l v a l u e s depending upon t h e o b s e r v a t i o n s ) . Our s i m u l a t i o n

s t u d y i n d i c a t e s t h a t a t l a r g e sample s i z e s they have b e t t e r power

t h a n T-WL f o r medium and heavy t a i l e d d i s t r i b u t i o n s . F o r s m a l l

sample s i z e s , t h e a b i l i t y t o o b t a i n e x a c t t e s t s by r e f e r r i n g t o

e x i s t i n g t a b l e s of t h e s i g n t e s t and t h e s igned-rank t e s t would

f a v o r T-EQ and T-SQR o v e r T-WL even f o r l i g h t - t a i l e d

d i s t r i b u t i o n s . Therefore , based on t h e s i m u l a t i o n s tudy and o t h e r

c o n s i d e r a t i o n s noted i n S e c t i o n 1, f o r l i g h t - t a i l e d d i s t r i b u t i o n s

we recommend T-EQ f o r s m a l l and moderate sample s i z e s (making

e x a c t t e s t s f e a s i b l e ) a n d T-WL f o r l a r g e sample s i z e s . We

recommend T-SQR f o r moderate t o heavy- ta i l ed d i s t r i b u t i o n s . I f a

s i n g l e s t a t i s t i c is t o be recommended f o r t e s t i n g Ho, t h e n we

would f a v o r T-SQR.

7. ACKNOWLEDGEMENTS

'Ihe a u t h o r s a r e g r a t e f u l t o t h e r e f e r e e f o r h e l p f u l comments.

REFEKENCES

Anscombe, F .J. (1952). Large-Sample Theory of S e q u e n t i a l Est imation. Proc. Cambridge Phi los . Soc., 4 8 , 600-607.

Hajek, J. and S i d a k , Z. (1967). l 'heory of Rank T e s t s . Academic P r e s s , New York.

K a l b f l e i s c h , J . D . and P r e n t i c e , R.L. (1973). Plarginal L ike l ihoods Based on Cox's Regression and L i f e b d e l . Biometr ika, 6 0 , 267-278.

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009

2056 POPOVICH AND RAO

Popov ich , E.A. ( 1 9 8 3 ) . h o n p a r a m e t r i c A n a l y s i s of B i v a r i a t e Censored Gata. '

D i s s e r t a t i o n . U n i v e r s i t y of F l o r i d a .

Ramberg, J . S . , and bchmei se r , B.W. ( 1 4 7 2 ) . An approx in ia t e h e t h o d t o r Generating Synmetr i c Random V a r i a b l e s . Commnica t ions of t h ? A s s o c i a t i o n f o r Computing h a c h i n e r y , i n c . , 15, 987-9913.

Rand les , R.h., a n d Ldolfe, L.A. (1979) . I n t r o d u c t i o n t o t h e Theory of Nonparamet r i c S t a t i s t i c s . Wi ley , New York.

S e r f l i n g , R.J. (198b) . Approx ima t ion Theorems of h t h e m a t i c a l S t a t i s t i c s . W i l e y , Iden York.

S p r o u l e , R.N. ( 1 9 7 4 ) . Asympto t i c P r o p e r t i e s of U - S t a t i s t i c s . T rans . Am. Y ~ t h e m a t i c a l Soc . , 1 9 9 , 55-61;.

l i o o l s o n , K.F., and Lachenbruch, P.H. (1980) . Hank T e s t s f o r Censored hatchec P a i r s . B i o m e t r i k a , 6 7 , 597-606.

Woolson, R.F. , a n d Lachenbruch, P.A. (1984) . C o r r e c t i o n s t o 'Bank T e s t s f o r Censored h t c h e d P a i r s . ' B i o m e t r i k a , 7 1 , 220.

Received AugunL, 19 t i 5 ; Revbed hluy, 1 9 t i 5 .

Recommended by L . 51. W e i , Gco/rgc Wa~lungLon UvhehbLty, WuhCurzgLo~, DC

Downloaded By: [Popovich, Edward] At: 12:40 24 November 2009