Trial had 4 treatments & data collected over multiple days.
When I run the following code (see below), results show a linear effect (trt x day interaction). What do I need to do, to determine which day the linear trend(s) are in?
|
Contrasts
|
|
Label
|
Num DF
|
Den DF
|
F Value
|
Pr > F
|
| |
|
|
|
|
| |
|
|
|
|
| |
|
|
|
|
|
LINEAR trtXday
|
1
|
38.8
|
14.09
|
0.0006
|
|
QUADRADIC trtXday
|
1
|
64.2
|
0.88
|
0.3519
|
|
CUBIC trtXday
|
1
|
77.9
|
0.01
|
0.9265
|
PROC MIXED;
CLASS TRT DAY ID;
MODEL BWkg = TRT|DAY/DDFM=KR;
REPEATED DAY/SUBJECT=ID TYPE = toep;
CONTRAST 'LINEAR trtXday' trt*day
1.549194 1.161894 0.774597 0.387297 -3.6E-20 -0.3872982 -0.7745967 -1.1618949 -1.5492
0.516398 0.387298 0.258199 0.129099 -1.2E-20 -0.1290994 -0.2581989 -0.3872983 -0.5164
-0.516398 -0.387298 -0.258199 -0.129099 1.2E-20 0.1290994 0.2581989 0.3872983 0.5164
-1.549194 -1.161894 -0.774597 -0.387297 3.6E-20 0.3872982 0.7745967 1.1618949 1.5492;
CONTRAST 'QUADRADIC trtXday' trt*day
0.531816 0.132954 -0.151947 -0.322888 -0.379869 -0.322888 -0.151947 0.132954 0.5318
-0.531816 -0.132954 0.151947 0.322888 0.379869 0.322888 0.151947 -0.132954 -0.5318
-0.531816 -0.132954 0.151947 0.322888 0.379869 0.322888 0.151947 -0.132954 -0.5318
0.531816 0.132954 -0.151947 -0.322888 -0.379869 -0.322888 -0.151947 0.132954 0.5318;
CONTRAST 'CUBIC trtXday' trt*day
0.444949 -0.2224746 -0.4131671 -0.2860388 1.7E-21 0.286039 0.413167 0.222475 -0.4449
-1.334847 0.6674238 1.2395013 0.8581164 -5.1E-21 -0.858117 -1.239501 -0.667425 1.3348
1.334847 -0.6674238 -1.2395013 -0.8581164 5.1E-21 0.858117 1.239501 0.667425 -1.3348
-0.444949 0.2224746 0.4131671 0.2860388 -1.7E-21 -0.286039 -0.413167 -0.222475 0.4449;
LSmeans trt day trt*day/DIFF ADJUST=SIMULATE (REPORT SEED=121211) cl adjdfe=row SLICE=(DAY trt);
RUN; Quit;
