Thanks Steve. TD21 ... View more

Thanks Steve. These are the codes I used for ANOVA without the fixed covariate effects. I tried four CVs (CS, CSH, AR(1), and UN), and CSH gives the best fit statistics. PROC Mixed DATA=WORK.NFldGrav_Final PLOTS(ONLY) = (ResidualPanel(Marginal)); CLASS Block Treat Point Year Time; MODEL NM = Treat|Year|Point  Time(Year) Treat*Time(Year) Point*Time(Year) RANDOM Block Block*Treat  Block*Treat*Point; REPEATED Time/SUBJECT=Point(Block*Treat*Year) TYPE = CSH; LSMEANS Treat*Year Treat*Year*Point Treat*Time(Year)/ADJUST=TUKEY SLICE=(Treat Year Point Time)CL; LSMEANS Treat*Point*Time(Year)/ADJUST=TUKEY SLICE=(Treat Year Point Time)CL; RUN; The approach you mentioned for analysis that includes the fixed covariate effects seems logical. I just would like to know if there's a way we could do to detect collinearity using this approach. While I was browsing for ideas to do the analysis with fixed covariates, I found that one solution is to do principal component analysis first  then MLR, or MLR only with automatic selection and variance inflation factors to detect collinearity. However, I don't know how to do PCA in SAS, much more with RCBD. I had experienced doing MLR in SAS, but not with data from RCBD experiment. So, my approach will be to do MLR by treatment and point with both dependent variables and covariates averaged over three blocks. To illustrate: Treatment 1, Point 1 PROC REG DATA = WORK.NField  PLOTS (ONLY) = (CP); STEPWISE: MODEL NM = BD TN TC CNR WC TMP/ SELECTION = STEPWISE; RUN; PROC REG DATA = WORK.NField PLOTS; FULL:  MODEL NM = BD TN TC CNR WC TMP / VIF; RUN; . . . Treatment 3, Point 2 PROC REG DATA = WORK.NField  PLOTS (ONLY) = (CP); STEPWISE: MODEL NM= BD TN TC CNR WC TMP/ SELECTION = STEPWISE; RUN; PROC REG DATA = WORK.NField PLOTS; FULL:  MODEL NM = BD TN TC CNR WC TMP / VIF; RUN; I didn't proceed with this idea, because I am not sure how much information will be lost by averaging over 3 blocks, if ever doing the analysis per treatment is a correct approach, in addition to eliminating time out of the equation. The bottomline is, I just would like to know, if some (eliminate collinear var.) or all of the fixed covariate effects significantly explain the variation in the dependent variable for each treatment at each point within each treatment. Thanks. ... View more

Hi Steve: I tried both 1 and 2, but I was not able to have a successful run due to the following: "Warning: obtaining minimum variance quadratic unbiased estimates as starting values for covariance parameters failed." What do I need to change to avoid this error? Thanks, TD21 ... View more

Thank you Steve. ... View more

Hi Steve: What are the equivalent PROC GLIMMIX codes for the following PROC Mixed Codes in order to get the estimates of the mean and standard error from the log (base 10) scale to the original scale? Thanks. 1.) PROC SORT DATA = WORK.LOGKsat_MID_TRK_Pre_Post; By Block Point Sample Depth LOC Harvest; RUN; ODS GRAPHICS ON; PROC MIXED DATA = WORK.LOGKsat_MID_TRK_PrePost plots(only)=(ResidualPanel(marginal))PLOTS (MAXPOINTS = 2000000); CLASS Block POINT SAMPLE DEPTH LOC Harvest; MODEL LG_Ksat=Harvest|LOC|DEPTH/DDFM = Satterthwaite RESIDUAL; RANDOM Block Block*Harvest BLOCK*LOC*Harvest; RANDOM Depth/SUBJECT = Point(Block*LOC*Harvest) TYPE = AR(1); LSMEANS Harvest Depth Harvest*Depth/PDIFF ADJUST = TUKEY SLICE=(Harvest Depth); LSMEANS Harvest*LOC*Depth/PDIFF ADJUST = TUKEY SLICE=(Harvest LOC Depth); RUN; ODS GRAPHICS OFF; 2.) PROC SORT DATA = Work.logKsat_Post2014; BY Block Treatment LOC Point Depth Sample; RUN; ODS GRAPHICS ON; PROC MIXED DATA =Work.logKsat_Post2014 plots(only)=(ResidualPanel(marginal)) PLOTS(MAXPOINTS= 2000000); CLASS Block Treatment LOC Point  Depth Sample; MODEL LG_Ksat = Treatment|Depth LOC(Treatment) Depth*LOC(Treatment) /RESIDUAL DDFM = Satterthwaite; RANDOM Block Block*Treatment  Block*Treatment*LOC  ; RANDOM Depth/SUBJECT = Point(Block*Treatment*LOC) TYPE = AR(1) ; LSMEANS Treatment Depth Treatment*Depth/PDIFF ADJUST = TUKEY slice=(treatment depth); LSMEANS  LOC*Depth(Treatment)/PDIFF ADJUST = TUKEY slice=(treatment depth LOC); RUN; ODS GRAPHICS OFF; Thank you. TD21 ... View more

Hi Steve: I followed your latest advice (add week, analyze with HPMIXED then port to GLIMMIX). I seemed to have successfully run HPMIXED, but not with GLIMMIX. The error was about PARMSDATA which I highlighted below. Any suggestion how to fix this error? Pardon the long post. I just wanted to give you as much background information as I can. Thanks. TD 55   PROC HPMIXED DATA = WORK.WTDWOR_Final2014; *plots(only)=(ResidualPanel(marginal) NOTE: Writing HTML Body file: sashtml1.htm 56                       StudentPanel(conditional)) PLOTS(MAXPOINTS= 2000000); 57       CLASS Block Treat Year Seas Week Day; 58       MODEL WTD = Treat|Year|Seas|Week; 59       RANDOM INT Treat/SUBJECT=Block; 60       REPEATED DAY/SUBJECT=Block*Treat*Year*Seas*Week type=ar(1) R; 61       TEST Treat|Year|Seas|Week; 62       ODS OUTPUT COVPARMS=HPMEstimate; 63   RUN; NOTE: 129 observations are excluded because of: missing response values (n=129). NOTE: Convergence criterion (GCONV=1E-8) satisfied. NOTE: The data set WORK.HPMESTIMATE has 4 observations and 3 variables. NOTE: PROCEDURE HPMIXED used (Total process time):       real time           2:32:48.21       cpu time            1:24:56.25 64 65   ODS GRAPHICS ON; 66   PROC GLIMMIX DATA = WORK.WTDWOR_Final2014; 67       CLASS Block Treat Year Seas Week Day; 68       MODEL WTD = Treat|Year|Seas|Week; 69       RANDOM INT Treat/SUBJECT=Block; 70       PARMS/PDATA=HPMEstimate hold=1,2 noiter; 71       LSMEANS Treat|Year|Seas|Week/ADJUST=Tukey; 72   RUN; NOTE: Some observations are not used in the analysis because of: missing response values (n=129). ERROR: The PARMSDATA= data set must have either 2 more or 1 less observations. NOTE: The SAS System stopped processing this step because of errors. NOTE: PROCEDURE GLIMMIX used (Total process time):       real time           40.35 seconds       cpu time            37.02 seconds ... View more

I have 92 days in 2009; 365 x 3 = 1095 days in 2010, 2011, and 2013; and 366 in 2012. So in each treatment I have 1553 days of data. By the way, I didn't include Julian date in my dataset but rather Julian Day. In short, I have a column for year, separate column for season, and day. For example in 2010, I have column/variable for year which is 2010, another column for season (winter, spring, summer, and fall) and another season for Day (1 to 365). Thanks again. ... View more

Thanks Steve. As you may have recognized, I am collecting the data every 15 minutes. I am both interested in main effects and simple effects (of treatments on WTD on annual basis, and on seasonal basis within a given year, if any). I average 240 samples of 15-minute WTD data to come up with an average for a day of a given season per treatment. The raw data are already on in terms of average per day and not the 15-minute interval. Analyzing by season (traditional 4 seasons) means averaging 120 days of WTD into one value per treatment. My point is, I could add a repeated statement in the model; however, the fact that I am averaging the data from 15-minute interval (240) to daily (120) to season makes the autocorrelation issue less of an issue. You are right with point number 2. I looked at the covariance parameters estimates and fit statistics (e.g. AIC values), the two models were very similar. I just would like to tell the model that DAY is nested within SEAS, and SEAS within Year and I am not sure if I am doing it right. Based on your comment about similarity in matrices, I think I am close. You mentioned about DAY. Any suggestion as to how to incorporate that into the model (DAY as nested into SEAS) as a code? Again, thanks for your time. ... View more

Thanks Rick... ... View more

Thanks Rick. ... View more