0530 sas實習課

26
SAS tutorial : GLM two-way anova & repeated measure 游游游

Upload: juchi-yu

Post on 27-Jul-2015

442 views

Category:

Documents


4 download

TRANSCRIPT

SAS tutorial : GLMtwo-way anova

& repeated measure

游如淇

PROC GLM ( 複習 )

PROC GLM 選項串 ;

CLASS 變相名稱串 ;

MODEL 依變項 = 實驗效果串 / 選項串 ;

MEANS 實驗效果串 / 選項串 ;

CONTRAST “ 標題” 效果名稱串 各組效果係數 / 選項串 TEST H= 效果項 E= 殘差項RUN;

• 還有 ESTIMATE、RANDOM、 LSMEAN… 等其他 statement ,跟 PROC ANOVA 差不多

Nominal or ordinal scale

Some statements ( 複習 )• CONTRAST :作對比分析

– CONTRAST “ 標題” 效果名稱串 各組效果係數 / 選項串 ;• “ 標題”: 20 字內,不能有“ ;”• 效果名稱: 要在 model 中出現過,界定所要比較的效果• 係數:總合為 0 以空格隔開

– E.g. CONTRAST ‘A1B1 vs A2B2’ A*B 1 0 0 -1;– E.g. CONTRAST ‘b at a1’ a*b 3 3 -2 -2 -2 0 0 0 0 0;

• 選項串: e, ss3……

• TEST :指定效果項與殘差項作 F 檢定– TEST H= 效果項 E= 殘差項 ;– E.g. TEST h=A*B e=S*A*B;

• CONTRAST 必須擺在 MODEL 之後, TEST、MANOVA、REPEATED、RANDOM 之前

• CLASS 需擺在 MODEL 之前

2-way ANOVA

Example 39.1title ’Balanced Data from Randomized Complete Block’;

data plants;

input Type $ @;

do Block = 1 to 3;

input StemLength @;

output;

end;

datalines;

Clarion 32.7 32.3 31.5

Clinton 32.1 29.7 29.1

Knox 35.7 35.9 33.1

O’Neill 36.0 34.2 31.2

Compost 31.8 28.0 29.2

Wabash 38.2 37.8 31.9

Webster 32.5 31.1 29.7

;

Example 39.1proc glm;

class Block Type;

model StemLength = Block Type;

run;

proc glm order=data;

class Block Type;

model StemLength = Block Type / solution;

/*----------------------------------clrn-cltn-knox-onel-cpst-wbsh-wstr */

contrast 'Compost vs. others' Type -1 -1 -1 -1 6 -1 -1;

contrast 'River soils vs. non' Type -1 -1 -1 -1 0 5 -1,

Type -1 4 -1 -1 0 0 -1;

contrast 'Glacial vs. drift' Type -1 0 1 1 0 0 -1;

contrast 'Clarion vs. Webster' Type -1 0 0 0 0 0 1;

contrast 'Knox vs. O’Neill' Type 0 0 1 -1 0 0 0;

run;

means Type / waller regwq;

run;

Result-1

Result-2

Result-3

Result-4

Repeated measure

Repeated measures

• 如果每個參與者都作了某些作業/回答某些題目。

• 那麼這些題目/作業便稱為重複量數( repeated measures )。– 或受試者內( within subject )設計。

• 可以去除來自個體差異的 variance 。

Example: 2-way repeated ANOVA

Source SS df MS FA A / (S*A)

B B / (S*B)

A*B A*B / (S*A*B)

S

S*A

S*B

S*A*B

Error

data aaa;

do S= 1 to 3;

do A=1 to 2;

do B=1 to 2;

input x @@;

output;

end;

end;

end;

datalines;

5 10 20 10

6 11 22 13

4 12 24 9

;

GLM for repeated measures

GLM for repeated measures: code

Between

proc glm data=AAA;

class A B S;

model x=A|B|S;

test h=A e=A*S;

test h=B e=B*S;

test h=A*B e=A*B*S;

mean A B A*B/tukey alpha=.5

run;

要考慮的變項( 分子 )

要比較的值

要知道的變項( 分母 )error

Result (between) : Interaction plot

Result (between)

Result (within)

DF 被 model 用完了,沒有 error 可以除?

Result

Between

Dummy coding

data k;input factor ncases @@;x0=1;x1=0;x2=0;x3=0;x4=0;if factor=1 then x1=1;if factor=2 then x2=1;if factor=3 then x3=1;if factor=4 then x4=1;cards;1 12 1 18 2 14 2 12 2 13 3 19 3 17 3 21 4 24 4 30;proc print;var factor ncases x0 x1 x2 x3 x4;

proc glm;class factor;model ncases=factor;run;proc reg;model ncases=x1 x2 x3 x4;run;proc reg;model ncases=x1 x2 x3;run;

Dummy coding

Dummy coding

Dummy coding – proc reg

Dummy coding – proc reg

Effect coding

data k;input factor ncases @@;x1=0;x2=0;x3=0;if factor=1 then x1=1;if factor=2 then x2=1;if factor=3 then x3=1;if factor=4 then do; x1=-1;x2=-1;x3=-1; endcards;1 12 1 18 2 14 2 12 2 13 3 19 3 17 3 21 4 24 4 30;

proc glm;class factor;model ncases=factor;run;proc reg;model ncases=x1 x2 x3 x4;run;proc reg;model ncases=x1 x2 x3;run;

2-way effect codingdata k;input area ses ncases @@;x1=0;y1=0;y2=0;if area=1 then x1=1;if area=2 then do; x1=-1; end;if ses=1 then y1=1;if ses=2 then y2=1;If ses=3 then do; y1=-1;y2=-1; end;x1y1=x1*y1x1y2=x1*y2cards;………………………………;

proc print;proc reg;model ncases = x1 y1 y2 x1y1 x1y2;run;