R: LS Means Analysis produces NAs?












0















I am running an linear model regression analysis script and I am running emmeans (ls means) on my model but I am getting a whole of NA's not sure why... Here is what I have run:



   setwd("C:/Users/wkmus/Desktop/R-Stuff")
### yeild-twt
ASM_Data<-read.csv("ASM_FIELD_18_SUMM_wm.csv",header=TRUE, na.strings=".")
head(ASM_Data)
str(ASM_Data)
####"NA" values in table are labeled as "." colored orange
ASM_Data$REP <- as.factor(ASM_Data$REP)
head(ASM_Data$REP)
ASM_Data$ENTRY_NO <-as.factor(ASM_Data$ENTRY_NO)
head(ASM_Data$ENTRY_NO)
ASM_Data$RANGE<-as.factor(ASM_Data$RANGE)
head(ASM_Data$RANGE)
ASM_Data$PLOT_ID<-as.factor(ASM_Data$PLOT_ID)
head(ASM_Data$PLOT_ID)
ASM_Data$PLOT<-as.factor(ASM_Data$PLOT)
head(ASM_Data$PLOT)
ASM_Data$ROW<-as.factor(ASM_Data$ROW)
head(ASM_Data$ROW)
ASM_Data$REP <- as.numeric(as.character(ASM_Data$REP))
head(ASM_Data$REP)
ASM_Data$TWT_g.li <- as.numeric(as.character(ASM_Data$TWT_g.li))
ASM_Data$Yield_kg.ha <- as.numeric(as.character(ASM_Data$Yield_kg.ha))
ASM_Data$PhysMat_Julian <- as.numeric(as.character(ASM_Data$PhysMat_Julian))
ASM_Data$flowering <- as.numeric(as.character(ASM_Data$flowering))
ASM_Data$height <- as.numeric(as.character(ASM_Data$height))
ASM_Data$CLEAN.WT <- as.numeric(as.character(ASM_Data$CLEAN.WT))
ASM_Data$GRAV.TEST.WEIGHT <-as.numeric(as.character(ASM_Data$GRAV.TEST.WEIGHT))
str(ASM_Data)

library(lme4)
#library(lsmeans)
library(emmeans)


Here is the data frame:



  > str(ASM_Data)
'data.frame': 270 obs. of 20 variables:
$ TRIAL_ID : Factor w/ 1 level "18ASM_OvOv": 1 1 1 1 1 1 1 1 1 1 ...
$ PLOT_ID : Factor w/ 270 levels "18ASM_OvOv_002",..: 1 2 3 4 5 6 7 8 9 10 ...
$ PLOT : Factor w/ 270 levels "2","3","4","5",..: 1 2 3 4 5 6 7 8 9 10 ...
$ ROW : Factor w/ 20 levels "1","2","3","4",..: 1 1 1 1 1 1 1 1 1 1 ...
$ RANGE : Factor w/ 15 levels "1","2","3","4",..: 2 3 4 5 6 7 8 9 10 12 ...
$ REP : num 1 1 1 1 1 1 1 1 1 1 ...
$ MP : int 1 1 1 1 1 1 1 1 1 1 ...
$ SUB.PLOT : Factor w/ 6 levels "A","B","C","D",..: 1 1 1 1 2 2 2 2 2 3 ...
$ ENTRY_NO : Factor w/ 139 levels "840","850","851",..: 116 82 87 134 77 120 34 62 48 136 ...
$ height : num 74 70 73 80 70 73 75 68 65 68 ...
$ flowering : num 133 133 134 134 133 131 133 137 134 132 ...
$ CLEAN.WT : num 1072 929 952 1149 1014 ...
$ GRAV.TEST.WEIGHT : num 349 309 332 340 325 ...
$ TWT_g.li : num 699 618 663 681 650 684 673 641 585 646 ...
$ Yield_kg.ha : num 2073 1797 1841 2222 1961 ...
$ Chaff.Color : Factor w/ 3 levels "Bronze","Mixed",..: 1 3 3 1 1 1 1 3 1 3 ...
$ CHAFF_COLOR_SCALE: int 2 1 1 2 2 2 2 1 2 1 ...
$ PhysMat : Factor w/ 3 levels "6/12/2018","6/13/2018",..: 1 1 1 1 1 1 1 1 1 1 ...
$ PhysMat_Julian : num 163 163 163 163 163 163 163 163 163 163 ...
$ PEDIGREE : Factor w/ 1 level "OVERLEY/OVERLAND": 1 1 1 1 1 1 1 1 1 1 ...


This is the head of ASM Data:



 head(ASM_Data)
`TRIAL_ID PLOT_ID PLOT ROW RANGE REP MP SUB.PLOT ENTRY_NO height flowering CLEAN.WT GRAV.TEST.WEIGHT TWT_g.li`
1 18ASM_OvOv 18ASM_OvOv_002 2 1 2 1 1 A 965 74 133 1071.5 349.37 699
2 18ASM_OvOv 18ASM_OvOv_003 3 1 3 1 1 A 931 70 133 928.8 309.13 618
3 18ASM_OvOv 18ASM_OvOv_004 4 1 4 1 1 A 936 73 134 951.8 331.70 663
4 18ASM_OvOv 18ASM_OvOv_005 5 1 5 1 1 A 983 80 134 1148.6 340.47 681
5 18ASM_OvOv 18ASM_OvOv_006 6 1 6 1 1 B 926 70 133 1014.0 324.95 650
6 18ASM_OvOv 18ASM_OvOv_007 7 1 7 1 1 B 969 73 131 1076.6 342.09 684
Yield_kg.ha Chaff.Color CHAFF_COLOR_SCALE PhysMat PhysMat_Julian PEDIGREE
1 2073 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
2 1797 White 1 6/12/2018 163 OVERLEY/OVERLAND
3 1841 White 1 6/12/2018 163 OVERLEY/OVERLAND
4 2222 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
5 1961 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
6 2082 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND


I am looking at a linear model dealing with test weight.



This is what I ran:



ASM_Data$TWT_g.li <- as.numeric(as.character((ASM_Data$TWT_g.li)))
head(ASM_Data$TWT_g.li)



ASM_YIELD_1 <- lm(TWT_g.li~ENTRY_NO + REP + SUB.BLOCK, data=ASM_Data)
anova(ASM_YIELD_1)
summary(ASM_YIELD_1)
emmeans(ASM_YIELD_1, "ENTRY_NO") ###########ADJ. MEANS


I get an output for anova



anova(ASM_YIELD_1)
Analysis of Variance Table

Response: TWT_g.li
Df Sum Sq Mean Sq F value Pr(>F)
ENTRY_NO 138 217949 1579 7.0339 < 2e-16 ***
REP 1 66410 66410 295.7683 < 2e-16 ***
SUB.BLOCK 4 1917 479 2.1348 0.08035 .
Residuals 125 28067 225
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1


but for emmeans I get something like this:



ENTRY_NO emmean SE df asymp.LCL asymp.UCL
840 nonEst NA NA NA NA
850 nonEst NA NA NA NA
851 nonEst NA NA NA NA
852 nonEst NA NA NA NA
853 nonEst NA NA NA NA
854 nonEst NA NA NA NA
855 nonEst NA NA NA NA
857 nonEst NA NA NA NA
858 nonEst NA NA NA NA
859 nonEst NA NA NA NA


I do have outliers in my data which is indicated by a "." in my data but that's the only thing I can think of which is off.



When I run with(ASM_Data, table(ENTRY_NO, REP, SUB.BLOCK))



this is what I have:



 with(ASM_Data, table(ENTRY_NO,REP,SUB.BLOCK))
, , SUB.BLOCK = A

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 0 0
854 0 0
855 0 0
857 0 0
858 0 0
859 0 0
860 0 0
861 0 0
862 0 0
863 1 0
864 0 0
865 1 0
866 1 0
867 0 0
868 0 0
869 1 0
870 1 0
871 0 0
872 0 0
873 0 0
874 0 0
875 0 0
876 0 0
877 0 0
878 0 0
879 1 0
880 0 0
881 0 0
882 0 0
883 0 0
884 0 0
885 1 0
886 0 0
887 1 0
888 1 0
889 1 0
890 0 0
891 1 0
892 0 0
893 0 0
894 0 0
895 0 0
896 1 0
897 0 0
898 0 0
899 0 0
900 1 0
901 1 0
902 0 0
903 0 0
904 1 0
905 1 0
906 0 0
907 1 0
908 1 0
909 0 0
910 0 0
911 0 0
912 0 0
913 0 0
914 0 0
915 0 0
916 1 0
917 0 0
918 0 0
919 1 0
920 0 0
921 0 0
922 0 0
923 1 0
924 0 0
925 0 0
926 0 0
927 1 0
928 0 0
929 0 0
930 0 0
931 1 0
932 0 0
933 0 0
934 0 0
935 0 0
936 1 0
937 0 0
938 1 0
939 1 0
940 0 0
941 1 0
942 0 0
943 1 0
944 0 0
945 0 0
946 0 0
947 0 0
948 1 0
949 0 0
950 1 0
951 0 0
952 0 0
953 0 0
954 0 0
955 1 0
956 1 0
957 1 0
958 1 0
959 0 0
960 0 0
961 0 0
962 0 0
963 0 0
964 0 0
965 1 0
966 0 0
967 1 0
968 0 0
969 0 0
970 1 0
971 0 0
972 0 0
973 0 0
974 1 0
975 0 0
976 0 0
977 0 0
978 1 0
979 0 0
980 0 0
981 0 0
982 0 0
983 1 0
984 1 0
985 0 0
986 1 0
987 3 0
988 0 0

, , SUB.BLOCK = B

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 1 0
854 0 0
855 0 0
857 0 0
858 0 0
859 0 0
860 0 0
861 1 0
862 0 0
863 0 0
864 0 0
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 1 0
872 0 0
873 0 0
874 0 0
875 0 0
876 1 0
877 1 0
878 1 0
879 0 0
880 1 0
881 0 0
882 1 0
883 1 0
884 1 0
885 0 0
886 0 0
887 0 0
888 0 0
889 0 0
890 1 0
891 0 0
892 1 0
893 1 0
894 1 0
895 1 0
896 0 0
897 1 0
898 0 0
899 0 0
900 0 0
901 0 0
902 1 0
903 0 0
904 0 0
905 0 0
906 0 0
907 0 0
908 0 0
909 1 0
910 0 0
911 1 0
912 0 0
913 1 0
914 0 0
915 0 0
916 0 0
917 0 0
918 0 0
919 0 0
920 1 0
921 1 0
922 0 0
923 0 0
924 0 0
925 1 0
926 1 0
927 0 0
928 0 0
929 0 0
930 1 0
931 0 0
932 1 0
933 0 0
934 1 0
935 0 0
936 0 0
937 1 0
938 0 0
939 0 0
940 1 0
941 0 0
942 0 0
943 0 0
944 0 0
945 1 0
946 0 0
947 1 0
948 0 0
949 0 0
950 0 0
951 1 0
952 0 0
953 0 0
954 1 0
955 0 0
956 0 0
957 0 0
958 0 0
959 1 0
960 0 0
961 0 0
962 1 0
963 0 0
964 0 0
965 0 0
966 0 0
967 0 0
968 0 0
969 1 0
970 0 0
971 0 0
972 0 0
973 0 0
974 0 0
975 0 0
976 1 0
977 1 0
978 0 0
979 0 0
980 0 0
981 1 0
982 1 0
983 0 0
984 0 0
985 3 0
986 0 0
987 1 0
988 1 0

, , SUB.BLOCK = C

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 0 0
854 0 0
855 0 0
857 1 0
858 0 0
859 1 0
860 0 0
861 0 0
862 1 0
863 0 0
864 0 0
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 0 0
872 1 0
873 0 0
874 0 0
875 0 0
876 0 0
877 0 0
878 0 0
879 0 0
880 0 0
881 1 0
882 0 0
883 0 0
884 0 0
885 0 0
886 1 0
887 0 0
888 0 0
889 0 0
890 0 0
891 0 0
892 0 0
893 0 0
894 0 0
895 0 0
896 0 0
897 0 0
898 1 0
899 1 0
900 0 0
901 0 0
902 0 0
903 1 0
904 0 0
905 0 0
906 1 0
907 0 0
908 0 0
909 0 0
910 1 0
911 0 0
912 1 0
913 0 0
914 1 0
915 1 0
916 0 0
917 1 0
918 1 0
919 0 0
920 0 0
921 0 0
922 1 0
923 0 0
924 1 0
925 0 0
926 0 0
927 0 0
928 1 0
929 1 0
930 0 0
931 0 0
932 0 0
933 1 0
934 0 0
935 1 0
936 0 0
937 0 0
938 0 0
939 0 0
940 0 0
941 0 0
942 1 0
943 0 0
944 1 0
945 0 0
946 1 0
947 0 0
948 0 0
949 1 0
950 0 0
951 0 0
952 1 0
953 1 0
954 0 0
955 0 0
956 0 0
957 0 0
958 0 0
959 0 0
960 1 0
961 1 0
962 0 0
963 1 0
964 1 0
965 0 0
966 1 0
967 0 0
968 1 0
969 0 0
970 0 0
971 1 0
972 1 0
973 1 0
974 0 0
975 1 0
976 0 0
977 0 0
978 1 0
979 2 0
980 0 0
981 0 0
982 0 0
983 0 0
984 0 0
985 1 0
986 3 0
987 0 0
988 0 0

, , SUB.BLOCK = D

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 1
853 0 0
854 0 0
855 0 0
857 0 0
858 0 1
859 0 0
860 0 1
861 0 0
862 0 0
863 0 0
864 0 1
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 0 0
872 0 0
873 0 0
874 0 0
875 0 1
876 0 0
877 0 0
878 0 1
879 0 0
880 0 1
881 0 1
882 0 1
883 0 1
884 0 1
885 0 0
886 0 0
887 0 0
888 0 0
889 0 0
890 0 0
891 0 0
892 0 1
893 0 0
894 0 0
895 0 0
896 0 0
897 0 1
898 0 0
899 0 1
900 0 0
901 0 0
902 0 1
903 0 0
904 0 0
905 0 0
906 0 0
907 0 0
908 0 0
909 0 0
910 0 0
911 0 0
912 0 0
913 0 1
914 0 1
915 0 1
916 0 0
917 0 1
918 0 1
919 0 0
920 0 0
921 0 1
922 0 1
923 0 0
924 0 0
925 0 0
926 0 0
927 0 0
928 0 0
929 0 1
930 0 1
931 0 0
932 0 0


Can someone please give me an idea of what is going wrong??



Thanks !










share|improve this question

























  • Can you show head(ASM_Data) in your question? It looks like ENTRY_NO may be an ID number, with only 1 observation per ENTRY_NO?

    – Marius
    Nov 20 '18 at 0:42











  • Hi~ let me add to the post.

    – wardah m
    Nov 20 '18 at 0:46











  • Where is the variable SUB.BLOCK? It’s in the model but not in the data.

    – rvl
    Nov 21 '18 at 4:47











  • Hi Sub.Plot is renamed as Sub.Block in my script.

    – wardah m
    Nov 21 '18 at 22:45
















0















I am running an linear model regression analysis script and I am running emmeans (ls means) on my model but I am getting a whole of NA's not sure why... Here is what I have run:



   setwd("C:/Users/wkmus/Desktop/R-Stuff")
### yeild-twt
ASM_Data<-read.csv("ASM_FIELD_18_SUMM_wm.csv",header=TRUE, na.strings=".")
head(ASM_Data)
str(ASM_Data)
####"NA" values in table are labeled as "." colored orange
ASM_Data$REP <- as.factor(ASM_Data$REP)
head(ASM_Data$REP)
ASM_Data$ENTRY_NO <-as.factor(ASM_Data$ENTRY_NO)
head(ASM_Data$ENTRY_NO)
ASM_Data$RANGE<-as.factor(ASM_Data$RANGE)
head(ASM_Data$RANGE)
ASM_Data$PLOT_ID<-as.factor(ASM_Data$PLOT_ID)
head(ASM_Data$PLOT_ID)
ASM_Data$PLOT<-as.factor(ASM_Data$PLOT)
head(ASM_Data$PLOT)
ASM_Data$ROW<-as.factor(ASM_Data$ROW)
head(ASM_Data$ROW)
ASM_Data$REP <- as.numeric(as.character(ASM_Data$REP))
head(ASM_Data$REP)
ASM_Data$TWT_g.li <- as.numeric(as.character(ASM_Data$TWT_g.li))
ASM_Data$Yield_kg.ha <- as.numeric(as.character(ASM_Data$Yield_kg.ha))
ASM_Data$PhysMat_Julian <- as.numeric(as.character(ASM_Data$PhysMat_Julian))
ASM_Data$flowering <- as.numeric(as.character(ASM_Data$flowering))
ASM_Data$height <- as.numeric(as.character(ASM_Data$height))
ASM_Data$CLEAN.WT <- as.numeric(as.character(ASM_Data$CLEAN.WT))
ASM_Data$GRAV.TEST.WEIGHT <-as.numeric(as.character(ASM_Data$GRAV.TEST.WEIGHT))
str(ASM_Data)

library(lme4)
#library(lsmeans)
library(emmeans)


Here is the data frame:



  > str(ASM_Data)
'data.frame': 270 obs. of 20 variables:
$ TRIAL_ID : Factor w/ 1 level "18ASM_OvOv": 1 1 1 1 1 1 1 1 1 1 ...
$ PLOT_ID : Factor w/ 270 levels "18ASM_OvOv_002",..: 1 2 3 4 5 6 7 8 9 10 ...
$ PLOT : Factor w/ 270 levels "2","3","4","5",..: 1 2 3 4 5 6 7 8 9 10 ...
$ ROW : Factor w/ 20 levels "1","2","3","4",..: 1 1 1 1 1 1 1 1 1 1 ...
$ RANGE : Factor w/ 15 levels "1","2","3","4",..: 2 3 4 5 6 7 8 9 10 12 ...
$ REP : num 1 1 1 1 1 1 1 1 1 1 ...
$ MP : int 1 1 1 1 1 1 1 1 1 1 ...
$ SUB.PLOT : Factor w/ 6 levels "A","B","C","D",..: 1 1 1 1 2 2 2 2 2 3 ...
$ ENTRY_NO : Factor w/ 139 levels "840","850","851",..: 116 82 87 134 77 120 34 62 48 136 ...
$ height : num 74 70 73 80 70 73 75 68 65 68 ...
$ flowering : num 133 133 134 134 133 131 133 137 134 132 ...
$ CLEAN.WT : num 1072 929 952 1149 1014 ...
$ GRAV.TEST.WEIGHT : num 349 309 332 340 325 ...
$ TWT_g.li : num 699 618 663 681 650 684 673 641 585 646 ...
$ Yield_kg.ha : num 2073 1797 1841 2222 1961 ...
$ Chaff.Color : Factor w/ 3 levels "Bronze","Mixed",..: 1 3 3 1 1 1 1 3 1 3 ...
$ CHAFF_COLOR_SCALE: int 2 1 1 2 2 2 2 1 2 1 ...
$ PhysMat : Factor w/ 3 levels "6/12/2018","6/13/2018",..: 1 1 1 1 1 1 1 1 1 1 ...
$ PhysMat_Julian : num 163 163 163 163 163 163 163 163 163 163 ...
$ PEDIGREE : Factor w/ 1 level "OVERLEY/OVERLAND": 1 1 1 1 1 1 1 1 1 1 ...


This is the head of ASM Data:



 head(ASM_Data)
`TRIAL_ID PLOT_ID PLOT ROW RANGE REP MP SUB.PLOT ENTRY_NO height flowering CLEAN.WT GRAV.TEST.WEIGHT TWT_g.li`
1 18ASM_OvOv 18ASM_OvOv_002 2 1 2 1 1 A 965 74 133 1071.5 349.37 699
2 18ASM_OvOv 18ASM_OvOv_003 3 1 3 1 1 A 931 70 133 928.8 309.13 618
3 18ASM_OvOv 18ASM_OvOv_004 4 1 4 1 1 A 936 73 134 951.8 331.70 663
4 18ASM_OvOv 18ASM_OvOv_005 5 1 5 1 1 A 983 80 134 1148.6 340.47 681
5 18ASM_OvOv 18ASM_OvOv_006 6 1 6 1 1 B 926 70 133 1014.0 324.95 650
6 18ASM_OvOv 18ASM_OvOv_007 7 1 7 1 1 B 969 73 131 1076.6 342.09 684
Yield_kg.ha Chaff.Color CHAFF_COLOR_SCALE PhysMat PhysMat_Julian PEDIGREE
1 2073 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
2 1797 White 1 6/12/2018 163 OVERLEY/OVERLAND
3 1841 White 1 6/12/2018 163 OVERLEY/OVERLAND
4 2222 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
5 1961 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
6 2082 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND


I am looking at a linear model dealing with test weight.



This is what I ran:



ASM_Data$TWT_g.li <- as.numeric(as.character((ASM_Data$TWT_g.li)))
head(ASM_Data$TWT_g.li)



ASM_YIELD_1 <- lm(TWT_g.li~ENTRY_NO + REP + SUB.BLOCK, data=ASM_Data)
anova(ASM_YIELD_1)
summary(ASM_YIELD_1)
emmeans(ASM_YIELD_1, "ENTRY_NO") ###########ADJ. MEANS


I get an output for anova



anova(ASM_YIELD_1)
Analysis of Variance Table

Response: TWT_g.li
Df Sum Sq Mean Sq F value Pr(>F)
ENTRY_NO 138 217949 1579 7.0339 < 2e-16 ***
REP 1 66410 66410 295.7683 < 2e-16 ***
SUB.BLOCK 4 1917 479 2.1348 0.08035 .
Residuals 125 28067 225
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1


but for emmeans I get something like this:



ENTRY_NO emmean SE df asymp.LCL asymp.UCL
840 nonEst NA NA NA NA
850 nonEst NA NA NA NA
851 nonEst NA NA NA NA
852 nonEst NA NA NA NA
853 nonEst NA NA NA NA
854 nonEst NA NA NA NA
855 nonEst NA NA NA NA
857 nonEst NA NA NA NA
858 nonEst NA NA NA NA
859 nonEst NA NA NA NA


I do have outliers in my data which is indicated by a "." in my data but that's the only thing I can think of which is off.



When I run with(ASM_Data, table(ENTRY_NO, REP, SUB.BLOCK))



this is what I have:



 with(ASM_Data, table(ENTRY_NO,REP,SUB.BLOCK))
, , SUB.BLOCK = A

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 0 0
854 0 0
855 0 0
857 0 0
858 0 0
859 0 0
860 0 0
861 0 0
862 0 0
863 1 0
864 0 0
865 1 0
866 1 0
867 0 0
868 0 0
869 1 0
870 1 0
871 0 0
872 0 0
873 0 0
874 0 0
875 0 0
876 0 0
877 0 0
878 0 0
879 1 0
880 0 0
881 0 0
882 0 0
883 0 0
884 0 0
885 1 0
886 0 0
887 1 0
888 1 0
889 1 0
890 0 0
891 1 0
892 0 0
893 0 0
894 0 0
895 0 0
896 1 0
897 0 0
898 0 0
899 0 0
900 1 0
901 1 0
902 0 0
903 0 0
904 1 0
905 1 0
906 0 0
907 1 0
908 1 0
909 0 0
910 0 0
911 0 0
912 0 0
913 0 0
914 0 0
915 0 0
916 1 0
917 0 0
918 0 0
919 1 0
920 0 0
921 0 0
922 0 0
923 1 0
924 0 0
925 0 0
926 0 0
927 1 0
928 0 0
929 0 0
930 0 0
931 1 0
932 0 0
933 0 0
934 0 0
935 0 0
936 1 0
937 0 0
938 1 0
939 1 0
940 0 0
941 1 0
942 0 0
943 1 0
944 0 0
945 0 0
946 0 0
947 0 0
948 1 0
949 0 0
950 1 0
951 0 0
952 0 0
953 0 0
954 0 0
955 1 0
956 1 0
957 1 0
958 1 0
959 0 0
960 0 0
961 0 0
962 0 0
963 0 0
964 0 0
965 1 0
966 0 0
967 1 0
968 0 0
969 0 0
970 1 0
971 0 0
972 0 0
973 0 0
974 1 0
975 0 0
976 0 0
977 0 0
978 1 0
979 0 0
980 0 0
981 0 0
982 0 0
983 1 0
984 1 0
985 0 0
986 1 0
987 3 0
988 0 0

, , SUB.BLOCK = B

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 1 0
854 0 0
855 0 0
857 0 0
858 0 0
859 0 0
860 0 0
861 1 0
862 0 0
863 0 0
864 0 0
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 1 0
872 0 0
873 0 0
874 0 0
875 0 0
876 1 0
877 1 0
878 1 0
879 0 0
880 1 0
881 0 0
882 1 0
883 1 0
884 1 0
885 0 0
886 0 0
887 0 0
888 0 0
889 0 0
890 1 0
891 0 0
892 1 0
893 1 0
894 1 0
895 1 0
896 0 0
897 1 0
898 0 0
899 0 0
900 0 0
901 0 0
902 1 0
903 0 0
904 0 0
905 0 0
906 0 0
907 0 0
908 0 0
909 1 0
910 0 0
911 1 0
912 0 0
913 1 0
914 0 0
915 0 0
916 0 0
917 0 0
918 0 0
919 0 0
920 1 0
921 1 0
922 0 0
923 0 0
924 0 0
925 1 0
926 1 0
927 0 0
928 0 0
929 0 0
930 1 0
931 0 0
932 1 0
933 0 0
934 1 0
935 0 0
936 0 0
937 1 0
938 0 0
939 0 0
940 1 0
941 0 0
942 0 0
943 0 0
944 0 0
945 1 0
946 0 0
947 1 0
948 0 0
949 0 0
950 0 0
951 1 0
952 0 0
953 0 0
954 1 0
955 0 0
956 0 0
957 0 0
958 0 0
959 1 0
960 0 0
961 0 0
962 1 0
963 0 0
964 0 0
965 0 0
966 0 0
967 0 0
968 0 0
969 1 0
970 0 0
971 0 0
972 0 0
973 0 0
974 0 0
975 0 0
976 1 0
977 1 0
978 0 0
979 0 0
980 0 0
981 1 0
982 1 0
983 0 0
984 0 0
985 3 0
986 0 0
987 1 0
988 1 0

, , SUB.BLOCK = C

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 0 0
854 0 0
855 0 0
857 1 0
858 0 0
859 1 0
860 0 0
861 0 0
862 1 0
863 0 0
864 0 0
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 0 0
872 1 0
873 0 0
874 0 0
875 0 0
876 0 0
877 0 0
878 0 0
879 0 0
880 0 0
881 1 0
882 0 0
883 0 0
884 0 0
885 0 0
886 1 0
887 0 0
888 0 0
889 0 0
890 0 0
891 0 0
892 0 0
893 0 0
894 0 0
895 0 0
896 0 0
897 0 0
898 1 0
899 1 0
900 0 0
901 0 0
902 0 0
903 1 0
904 0 0
905 0 0
906 1 0
907 0 0
908 0 0
909 0 0
910 1 0
911 0 0
912 1 0
913 0 0
914 1 0
915 1 0
916 0 0
917 1 0
918 1 0
919 0 0
920 0 0
921 0 0
922 1 0
923 0 0
924 1 0
925 0 0
926 0 0
927 0 0
928 1 0
929 1 0
930 0 0
931 0 0
932 0 0
933 1 0
934 0 0
935 1 0
936 0 0
937 0 0
938 0 0
939 0 0
940 0 0
941 0 0
942 1 0
943 0 0
944 1 0
945 0 0
946 1 0
947 0 0
948 0 0
949 1 0
950 0 0
951 0 0
952 1 0
953 1 0
954 0 0
955 0 0
956 0 0
957 0 0
958 0 0
959 0 0
960 1 0
961 1 0
962 0 0
963 1 0
964 1 0
965 0 0
966 1 0
967 0 0
968 1 0
969 0 0
970 0 0
971 1 0
972 1 0
973 1 0
974 0 0
975 1 0
976 0 0
977 0 0
978 1 0
979 2 0
980 0 0
981 0 0
982 0 0
983 0 0
984 0 0
985 1 0
986 3 0
987 0 0
988 0 0

, , SUB.BLOCK = D

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 1
853 0 0
854 0 0
855 0 0
857 0 0
858 0 1
859 0 0
860 0 1
861 0 0
862 0 0
863 0 0
864 0 1
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 0 0
872 0 0
873 0 0
874 0 0
875 0 1
876 0 0
877 0 0
878 0 1
879 0 0
880 0 1
881 0 1
882 0 1
883 0 1
884 0 1
885 0 0
886 0 0
887 0 0
888 0 0
889 0 0
890 0 0
891 0 0
892 0 1
893 0 0
894 0 0
895 0 0
896 0 0
897 0 1
898 0 0
899 0 1
900 0 0
901 0 0
902 0 1
903 0 0
904 0 0
905 0 0
906 0 0
907 0 0
908 0 0
909 0 0
910 0 0
911 0 0
912 0 0
913 0 1
914 0 1
915 0 1
916 0 0
917 0 1
918 0 1
919 0 0
920 0 0
921 0 1
922 0 1
923 0 0
924 0 0
925 0 0
926 0 0
927 0 0
928 0 0
929 0 1
930 0 1
931 0 0
932 0 0


Can someone please give me an idea of what is going wrong??



Thanks !










share|improve this question

























  • Can you show head(ASM_Data) in your question? It looks like ENTRY_NO may be an ID number, with only 1 observation per ENTRY_NO?

    – Marius
    Nov 20 '18 at 0:42











  • Hi~ let me add to the post.

    – wardah m
    Nov 20 '18 at 0:46











  • Where is the variable SUB.BLOCK? It’s in the model but not in the data.

    – rvl
    Nov 21 '18 at 4:47











  • Hi Sub.Plot is renamed as Sub.Block in my script.

    – wardah m
    Nov 21 '18 at 22:45














0












0








0








I am running an linear model regression analysis script and I am running emmeans (ls means) on my model but I am getting a whole of NA's not sure why... Here is what I have run:



   setwd("C:/Users/wkmus/Desktop/R-Stuff")
### yeild-twt
ASM_Data<-read.csv("ASM_FIELD_18_SUMM_wm.csv",header=TRUE, na.strings=".")
head(ASM_Data)
str(ASM_Data)
####"NA" values in table are labeled as "." colored orange
ASM_Data$REP <- as.factor(ASM_Data$REP)
head(ASM_Data$REP)
ASM_Data$ENTRY_NO <-as.factor(ASM_Data$ENTRY_NO)
head(ASM_Data$ENTRY_NO)
ASM_Data$RANGE<-as.factor(ASM_Data$RANGE)
head(ASM_Data$RANGE)
ASM_Data$PLOT_ID<-as.factor(ASM_Data$PLOT_ID)
head(ASM_Data$PLOT_ID)
ASM_Data$PLOT<-as.factor(ASM_Data$PLOT)
head(ASM_Data$PLOT)
ASM_Data$ROW<-as.factor(ASM_Data$ROW)
head(ASM_Data$ROW)
ASM_Data$REP <- as.numeric(as.character(ASM_Data$REP))
head(ASM_Data$REP)
ASM_Data$TWT_g.li <- as.numeric(as.character(ASM_Data$TWT_g.li))
ASM_Data$Yield_kg.ha <- as.numeric(as.character(ASM_Data$Yield_kg.ha))
ASM_Data$PhysMat_Julian <- as.numeric(as.character(ASM_Data$PhysMat_Julian))
ASM_Data$flowering <- as.numeric(as.character(ASM_Data$flowering))
ASM_Data$height <- as.numeric(as.character(ASM_Data$height))
ASM_Data$CLEAN.WT <- as.numeric(as.character(ASM_Data$CLEAN.WT))
ASM_Data$GRAV.TEST.WEIGHT <-as.numeric(as.character(ASM_Data$GRAV.TEST.WEIGHT))
str(ASM_Data)

library(lme4)
#library(lsmeans)
library(emmeans)


Here is the data frame:



  > str(ASM_Data)
'data.frame': 270 obs. of 20 variables:
$ TRIAL_ID : Factor w/ 1 level "18ASM_OvOv": 1 1 1 1 1 1 1 1 1 1 ...
$ PLOT_ID : Factor w/ 270 levels "18ASM_OvOv_002",..: 1 2 3 4 5 6 7 8 9 10 ...
$ PLOT : Factor w/ 270 levels "2","3","4","5",..: 1 2 3 4 5 6 7 8 9 10 ...
$ ROW : Factor w/ 20 levels "1","2","3","4",..: 1 1 1 1 1 1 1 1 1 1 ...
$ RANGE : Factor w/ 15 levels "1","2","3","4",..: 2 3 4 5 6 7 8 9 10 12 ...
$ REP : num 1 1 1 1 1 1 1 1 1 1 ...
$ MP : int 1 1 1 1 1 1 1 1 1 1 ...
$ SUB.PLOT : Factor w/ 6 levels "A","B","C","D",..: 1 1 1 1 2 2 2 2 2 3 ...
$ ENTRY_NO : Factor w/ 139 levels "840","850","851",..: 116 82 87 134 77 120 34 62 48 136 ...
$ height : num 74 70 73 80 70 73 75 68 65 68 ...
$ flowering : num 133 133 134 134 133 131 133 137 134 132 ...
$ CLEAN.WT : num 1072 929 952 1149 1014 ...
$ GRAV.TEST.WEIGHT : num 349 309 332 340 325 ...
$ TWT_g.li : num 699 618 663 681 650 684 673 641 585 646 ...
$ Yield_kg.ha : num 2073 1797 1841 2222 1961 ...
$ Chaff.Color : Factor w/ 3 levels "Bronze","Mixed",..: 1 3 3 1 1 1 1 3 1 3 ...
$ CHAFF_COLOR_SCALE: int 2 1 1 2 2 2 2 1 2 1 ...
$ PhysMat : Factor w/ 3 levels "6/12/2018","6/13/2018",..: 1 1 1 1 1 1 1 1 1 1 ...
$ PhysMat_Julian : num 163 163 163 163 163 163 163 163 163 163 ...
$ PEDIGREE : Factor w/ 1 level "OVERLEY/OVERLAND": 1 1 1 1 1 1 1 1 1 1 ...


This is the head of ASM Data:



 head(ASM_Data)
`TRIAL_ID PLOT_ID PLOT ROW RANGE REP MP SUB.PLOT ENTRY_NO height flowering CLEAN.WT GRAV.TEST.WEIGHT TWT_g.li`
1 18ASM_OvOv 18ASM_OvOv_002 2 1 2 1 1 A 965 74 133 1071.5 349.37 699
2 18ASM_OvOv 18ASM_OvOv_003 3 1 3 1 1 A 931 70 133 928.8 309.13 618
3 18ASM_OvOv 18ASM_OvOv_004 4 1 4 1 1 A 936 73 134 951.8 331.70 663
4 18ASM_OvOv 18ASM_OvOv_005 5 1 5 1 1 A 983 80 134 1148.6 340.47 681
5 18ASM_OvOv 18ASM_OvOv_006 6 1 6 1 1 B 926 70 133 1014.0 324.95 650
6 18ASM_OvOv 18ASM_OvOv_007 7 1 7 1 1 B 969 73 131 1076.6 342.09 684
Yield_kg.ha Chaff.Color CHAFF_COLOR_SCALE PhysMat PhysMat_Julian PEDIGREE
1 2073 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
2 1797 White 1 6/12/2018 163 OVERLEY/OVERLAND
3 1841 White 1 6/12/2018 163 OVERLEY/OVERLAND
4 2222 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
5 1961 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
6 2082 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND


I am looking at a linear model dealing with test weight.



This is what I ran:



ASM_Data$TWT_g.li <- as.numeric(as.character((ASM_Data$TWT_g.li)))
head(ASM_Data$TWT_g.li)



ASM_YIELD_1 <- lm(TWT_g.li~ENTRY_NO + REP + SUB.BLOCK, data=ASM_Data)
anova(ASM_YIELD_1)
summary(ASM_YIELD_1)
emmeans(ASM_YIELD_1, "ENTRY_NO") ###########ADJ. MEANS


I get an output for anova



anova(ASM_YIELD_1)
Analysis of Variance Table

Response: TWT_g.li
Df Sum Sq Mean Sq F value Pr(>F)
ENTRY_NO 138 217949 1579 7.0339 < 2e-16 ***
REP 1 66410 66410 295.7683 < 2e-16 ***
SUB.BLOCK 4 1917 479 2.1348 0.08035 .
Residuals 125 28067 225
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1


but for emmeans I get something like this:



ENTRY_NO emmean SE df asymp.LCL asymp.UCL
840 nonEst NA NA NA NA
850 nonEst NA NA NA NA
851 nonEst NA NA NA NA
852 nonEst NA NA NA NA
853 nonEst NA NA NA NA
854 nonEst NA NA NA NA
855 nonEst NA NA NA NA
857 nonEst NA NA NA NA
858 nonEst NA NA NA NA
859 nonEst NA NA NA NA


I do have outliers in my data which is indicated by a "." in my data but that's the only thing I can think of which is off.



When I run with(ASM_Data, table(ENTRY_NO, REP, SUB.BLOCK))



this is what I have:



 with(ASM_Data, table(ENTRY_NO,REP,SUB.BLOCK))
, , SUB.BLOCK = A

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 0 0
854 0 0
855 0 0
857 0 0
858 0 0
859 0 0
860 0 0
861 0 0
862 0 0
863 1 0
864 0 0
865 1 0
866 1 0
867 0 0
868 0 0
869 1 0
870 1 0
871 0 0
872 0 0
873 0 0
874 0 0
875 0 0
876 0 0
877 0 0
878 0 0
879 1 0
880 0 0
881 0 0
882 0 0
883 0 0
884 0 0
885 1 0
886 0 0
887 1 0
888 1 0
889 1 0
890 0 0
891 1 0
892 0 0
893 0 0
894 0 0
895 0 0
896 1 0
897 0 0
898 0 0
899 0 0
900 1 0
901 1 0
902 0 0
903 0 0
904 1 0
905 1 0
906 0 0
907 1 0
908 1 0
909 0 0
910 0 0
911 0 0
912 0 0
913 0 0
914 0 0
915 0 0
916 1 0
917 0 0
918 0 0
919 1 0
920 0 0
921 0 0
922 0 0
923 1 0
924 0 0
925 0 0
926 0 0
927 1 0
928 0 0
929 0 0
930 0 0
931 1 0
932 0 0
933 0 0
934 0 0
935 0 0
936 1 0
937 0 0
938 1 0
939 1 0
940 0 0
941 1 0
942 0 0
943 1 0
944 0 0
945 0 0
946 0 0
947 0 0
948 1 0
949 0 0
950 1 0
951 0 0
952 0 0
953 0 0
954 0 0
955 1 0
956 1 0
957 1 0
958 1 0
959 0 0
960 0 0
961 0 0
962 0 0
963 0 0
964 0 0
965 1 0
966 0 0
967 1 0
968 0 0
969 0 0
970 1 0
971 0 0
972 0 0
973 0 0
974 1 0
975 0 0
976 0 0
977 0 0
978 1 0
979 0 0
980 0 0
981 0 0
982 0 0
983 1 0
984 1 0
985 0 0
986 1 0
987 3 0
988 0 0

, , SUB.BLOCK = B

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 1 0
854 0 0
855 0 0
857 0 0
858 0 0
859 0 0
860 0 0
861 1 0
862 0 0
863 0 0
864 0 0
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 1 0
872 0 0
873 0 0
874 0 0
875 0 0
876 1 0
877 1 0
878 1 0
879 0 0
880 1 0
881 0 0
882 1 0
883 1 0
884 1 0
885 0 0
886 0 0
887 0 0
888 0 0
889 0 0
890 1 0
891 0 0
892 1 0
893 1 0
894 1 0
895 1 0
896 0 0
897 1 0
898 0 0
899 0 0
900 0 0
901 0 0
902 1 0
903 0 0
904 0 0
905 0 0
906 0 0
907 0 0
908 0 0
909 1 0
910 0 0
911 1 0
912 0 0
913 1 0
914 0 0
915 0 0
916 0 0
917 0 0
918 0 0
919 0 0
920 1 0
921 1 0
922 0 0
923 0 0
924 0 0
925 1 0
926 1 0
927 0 0
928 0 0
929 0 0
930 1 0
931 0 0
932 1 0
933 0 0
934 1 0
935 0 0
936 0 0
937 1 0
938 0 0
939 0 0
940 1 0
941 0 0
942 0 0
943 0 0
944 0 0
945 1 0
946 0 0
947 1 0
948 0 0
949 0 0
950 0 0
951 1 0
952 0 0
953 0 0
954 1 0
955 0 0
956 0 0
957 0 0
958 0 0
959 1 0
960 0 0
961 0 0
962 1 0
963 0 0
964 0 0
965 0 0
966 0 0
967 0 0
968 0 0
969 1 0
970 0 0
971 0 0
972 0 0
973 0 0
974 0 0
975 0 0
976 1 0
977 1 0
978 0 0
979 0 0
980 0 0
981 1 0
982 1 0
983 0 0
984 0 0
985 3 0
986 0 0
987 1 0
988 1 0

, , SUB.BLOCK = C

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 0 0
854 0 0
855 0 0
857 1 0
858 0 0
859 1 0
860 0 0
861 0 0
862 1 0
863 0 0
864 0 0
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 0 0
872 1 0
873 0 0
874 0 0
875 0 0
876 0 0
877 0 0
878 0 0
879 0 0
880 0 0
881 1 0
882 0 0
883 0 0
884 0 0
885 0 0
886 1 0
887 0 0
888 0 0
889 0 0
890 0 0
891 0 0
892 0 0
893 0 0
894 0 0
895 0 0
896 0 0
897 0 0
898 1 0
899 1 0
900 0 0
901 0 0
902 0 0
903 1 0
904 0 0
905 0 0
906 1 0
907 0 0
908 0 0
909 0 0
910 1 0
911 0 0
912 1 0
913 0 0
914 1 0
915 1 0
916 0 0
917 1 0
918 1 0
919 0 0
920 0 0
921 0 0
922 1 0
923 0 0
924 1 0
925 0 0
926 0 0
927 0 0
928 1 0
929 1 0
930 0 0
931 0 0
932 0 0
933 1 0
934 0 0
935 1 0
936 0 0
937 0 0
938 0 0
939 0 0
940 0 0
941 0 0
942 1 0
943 0 0
944 1 0
945 0 0
946 1 0
947 0 0
948 0 0
949 1 0
950 0 0
951 0 0
952 1 0
953 1 0
954 0 0
955 0 0
956 0 0
957 0 0
958 0 0
959 0 0
960 1 0
961 1 0
962 0 0
963 1 0
964 1 0
965 0 0
966 1 0
967 0 0
968 1 0
969 0 0
970 0 0
971 1 0
972 1 0
973 1 0
974 0 0
975 1 0
976 0 0
977 0 0
978 1 0
979 2 0
980 0 0
981 0 0
982 0 0
983 0 0
984 0 0
985 1 0
986 3 0
987 0 0
988 0 0

, , SUB.BLOCK = D

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 1
853 0 0
854 0 0
855 0 0
857 0 0
858 0 1
859 0 0
860 0 1
861 0 0
862 0 0
863 0 0
864 0 1
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 0 0
872 0 0
873 0 0
874 0 0
875 0 1
876 0 0
877 0 0
878 0 1
879 0 0
880 0 1
881 0 1
882 0 1
883 0 1
884 0 1
885 0 0
886 0 0
887 0 0
888 0 0
889 0 0
890 0 0
891 0 0
892 0 1
893 0 0
894 0 0
895 0 0
896 0 0
897 0 1
898 0 0
899 0 1
900 0 0
901 0 0
902 0 1
903 0 0
904 0 0
905 0 0
906 0 0
907 0 0
908 0 0
909 0 0
910 0 0
911 0 0
912 0 0
913 0 1
914 0 1
915 0 1
916 0 0
917 0 1
918 0 1
919 0 0
920 0 0
921 0 1
922 0 1
923 0 0
924 0 0
925 0 0
926 0 0
927 0 0
928 0 0
929 0 1
930 0 1
931 0 0
932 0 0


Can someone please give me an idea of what is going wrong??



Thanks !










share|improve this question
















I am running an linear model regression analysis script and I am running emmeans (ls means) on my model but I am getting a whole of NA's not sure why... Here is what I have run:



   setwd("C:/Users/wkmus/Desktop/R-Stuff")
### yeild-twt
ASM_Data<-read.csv("ASM_FIELD_18_SUMM_wm.csv",header=TRUE, na.strings=".")
head(ASM_Data)
str(ASM_Data)
####"NA" values in table are labeled as "." colored orange
ASM_Data$REP <- as.factor(ASM_Data$REP)
head(ASM_Data$REP)
ASM_Data$ENTRY_NO <-as.factor(ASM_Data$ENTRY_NO)
head(ASM_Data$ENTRY_NO)
ASM_Data$RANGE<-as.factor(ASM_Data$RANGE)
head(ASM_Data$RANGE)
ASM_Data$PLOT_ID<-as.factor(ASM_Data$PLOT_ID)
head(ASM_Data$PLOT_ID)
ASM_Data$PLOT<-as.factor(ASM_Data$PLOT)
head(ASM_Data$PLOT)
ASM_Data$ROW<-as.factor(ASM_Data$ROW)
head(ASM_Data$ROW)
ASM_Data$REP <- as.numeric(as.character(ASM_Data$REP))
head(ASM_Data$REP)
ASM_Data$TWT_g.li <- as.numeric(as.character(ASM_Data$TWT_g.li))
ASM_Data$Yield_kg.ha <- as.numeric(as.character(ASM_Data$Yield_kg.ha))
ASM_Data$PhysMat_Julian <- as.numeric(as.character(ASM_Data$PhysMat_Julian))
ASM_Data$flowering <- as.numeric(as.character(ASM_Data$flowering))
ASM_Data$height <- as.numeric(as.character(ASM_Data$height))
ASM_Data$CLEAN.WT <- as.numeric(as.character(ASM_Data$CLEAN.WT))
ASM_Data$GRAV.TEST.WEIGHT <-as.numeric(as.character(ASM_Data$GRAV.TEST.WEIGHT))
str(ASM_Data)

library(lme4)
#library(lsmeans)
library(emmeans)


Here is the data frame:



  > str(ASM_Data)
'data.frame': 270 obs. of 20 variables:
$ TRIAL_ID : Factor w/ 1 level "18ASM_OvOv": 1 1 1 1 1 1 1 1 1 1 ...
$ PLOT_ID : Factor w/ 270 levels "18ASM_OvOv_002",..: 1 2 3 4 5 6 7 8 9 10 ...
$ PLOT : Factor w/ 270 levels "2","3","4","5",..: 1 2 3 4 5 6 7 8 9 10 ...
$ ROW : Factor w/ 20 levels "1","2","3","4",..: 1 1 1 1 1 1 1 1 1 1 ...
$ RANGE : Factor w/ 15 levels "1","2","3","4",..: 2 3 4 5 6 7 8 9 10 12 ...
$ REP : num 1 1 1 1 1 1 1 1 1 1 ...
$ MP : int 1 1 1 1 1 1 1 1 1 1 ...
$ SUB.PLOT : Factor w/ 6 levels "A","B","C","D",..: 1 1 1 1 2 2 2 2 2 3 ...
$ ENTRY_NO : Factor w/ 139 levels "840","850","851",..: 116 82 87 134 77 120 34 62 48 136 ...
$ height : num 74 70 73 80 70 73 75 68 65 68 ...
$ flowering : num 133 133 134 134 133 131 133 137 134 132 ...
$ CLEAN.WT : num 1072 929 952 1149 1014 ...
$ GRAV.TEST.WEIGHT : num 349 309 332 340 325 ...
$ TWT_g.li : num 699 618 663 681 650 684 673 641 585 646 ...
$ Yield_kg.ha : num 2073 1797 1841 2222 1961 ...
$ Chaff.Color : Factor w/ 3 levels "Bronze","Mixed",..: 1 3 3 1 1 1 1 3 1 3 ...
$ CHAFF_COLOR_SCALE: int 2 1 1 2 2 2 2 1 2 1 ...
$ PhysMat : Factor w/ 3 levels "6/12/2018","6/13/2018",..: 1 1 1 1 1 1 1 1 1 1 ...
$ PhysMat_Julian : num 163 163 163 163 163 163 163 163 163 163 ...
$ PEDIGREE : Factor w/ 1 level "OVERLEY/OVERLAND": 1 1 1 1 1 1 1 1 1 1 ...


This is the head of ASM Data:



 head(ASM_Data)
`TRIAL_ID PLOT_ID PLOT ROW RANGE REP MP SUB.PLOT ENTRY_NO height flowering CLEAN.WT GRAV.TEST.WEIGHT TWT_g.li`
1 18ASM_OvOv 18ASM_OvOv_002 2 1 2 1 1 A 965 74 133 1071.5 349.37 699
2 18ASM_OvOv 18ASM_OvOv_003 3 1 3 1 1 A 931 70 133 928.8 309.13 618
3 18ASM_OvOv 18ASM_OvOv_004 4 1 4 1 1 A 936 73 134 951.8 331.70 663
4 18ASM_OvOv 18ASM_OvOv_005 5 1 5 1 1 A 983 80 134 1148.6 340.47 681
5 18ASM_OvOv 18ASM_OvOv_006 6 1 6 1 1 B 926 70 133 1014.0 324.95 650
6 18ASM_OvOv 18ASM_OvOv_007 7 1 7 1 1 B 969 73 131 1076.6 342.09 684
Yield_kg.ha Chaff.Color CHAFF_COLOR_SCALE PhysMat PhysMat_Julian PEDIGREE
1 2073 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
2 1797 White 1 6/12/2018 163 OVERLEY/OVERLAND
3 1841 White 1 6/12/2018 163 OVERLEY/OVERLAND
4 2222 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
5 1961 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND
6 2082 Bronze 2 6/12/2018 163 OVERLEY/OVERLAND


I am looking at a linear model dealing with test weight.



This is what I ran:



ASM_Data$TWT_g.li <- as.numeric(as.character((ASM_Data$TWT_g.li)))
head(ASM_Data$TWT_g.li)



ASM_YIELD_1 <- lm(TWT_g.li~ENTRY_NO + REP + SUB.BLOCK, data=ASM_Data)
anova(ASM_YIELD_1)
summary(ASM_YIELD_1)
emmeans(ASM_YIELD_1, "ENTRY_NO") ###########ADJ. MEANS


I get an output for anova



anova(ASM_YIELD_1)
Analysis of Variance Table

Response: TWT_g.li
Df Sum Sq Mean Sq F value Pr(>F)
ENTRY_NO 138 217949 1579 7.0339 < 2e-16 ***
REP 1 66410 66410 295.7683 < 2e-16 ***
SUB.BLOCK 4 1917 479 2.1348 0.08035 .
Residuals 125 28067 225
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1


but for emmeans I get something like this:



ENTRY_NO emmean SE df asymp.LCL asymp.UCL
840 nonEst NA NA NA NA
850 nonEst NA NA NA NA
851 nonEst NA NA NA NA
852 nonEst NA NA NA NA
853 nonEst NA NA NA NA
854 nonEst NA NA NA NA
855 nonEst NA NA NA NA
857 nonEst NA NA NA NA
858 nonEst NA NA NA NA
859 nonEst NA NA NA NA


I do have outliers in my data which is indicated by a "." in my data but that's the only thing I can think of which is off.



When I run with(ASM_Data, table(ENTRY_NO, REP, SUB.BLOCK))



this is what I have:



 with(ASM_Data, table(ENTRY_NO,REP,SUB.BLOCK))
, , SUB.BLOCK = A

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 0 0
854 0 0
855 0 0
857 0 0
858 0 0
859 0 0
860 0 0
861 0 0
862 0 0
863 1 0
864 0 0
865 1 0
866 1 0
867 0 0
868 0 0
869 1 0
870 1 0
871 0 0
872 0 0
873 0 0
874 0 0
875 0 0
876 0 0
877 0 0
878 0 0
879 1 0
880 0 0
881 0 0
882 0 0
883 0 0
884 0 0
885 1 0
886 0 0
887 1 0
888 1 0
889 1 0
890 0 0
891 1 0
892 0 0
893 0 0
894 0 0
895 0 0
896 1 0
897 0 0
898 0 0
899 0 0
900 1 0
901 1 0
902 0 0
903 0 0
904 1 0
905 1 0
906 0 0
907 1 0
908 1 0
909 0 0
910 0 0
911 0 0
912 0 0
913 0 0
914 0 0
915 0 0
916 1 0
917 0 0
918 0 0
919 1 0
920 0 0
921 0 0
922 0 0
923 1 0
924 0 0
925 0 0
926 0 0
927 1 0
928 0 0
929 0 0
930 0 0
931 1 0
932 0 0
933 0 0
934 0 0
935 0 0
936 1 0
937 0 0
938 1 0
939 1 0
940 0 0
941 1 0
942 0 0
943 1 0
944 0 0
945 0 0
946 0 0
947 0 0
948 1 0
949 0 0
950 1 0
951 0 0
952 0 0
953 0 0
954 0 0
955 1 0
956 1 0
957 1 0
958 1 0
959 0 0
960 0 0
961 0 0
962 0 0
963 0 0
964 0 0
965 1 0
966 0 0
967 1 0
968 0 0
969 0 0
970 1 0
971 0 0
972 0 0
973 0 0
974 1 0
975 0 0
976 0 0
977 0 0
978 1 0
979 0 0
980 0 0
981 0 0
982 0 0
983 1 0
984 1 0
985 0 0
986 1 0
987 3 0
988 0 0

, , SUB.BLOCK = B

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 1 0
854 0 0
855 0 0
857 0 0
858 0 0
859 0 0
860 0 0
861 1 0
862 0 0
863 0 0
864 0 0
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 1 0
872 0 0
873 0 0
874 0 0
875 0 0
876 1 0
877 1 0
878 1 0
879 0 0
880 1 0
881 0 0
882 1 0
883 1 0
884 1 0
885 0 0
886 0 0
887 0 0
888 0 0
889 0 0
890 1 0
891 0 0
892 1 0
893 1 0
894 1 0
895 1 0
896 0 0
897 1 0
898 0 0
899 0 0
900 0 0
901 0 0
902 1 0
903 0 0
904 0 0
905 0 0
906 0 0
907 0 0
908 0 0
909 1 0
910 0 0
911 1 0
912 0 0
913 1 0
914 0 0
915 0 0
916 0 0
917 0 0
918 0 0
919 0 0
920 1 0
921 1 0
922 0 0
923 0 0
924 0 0
925 1 0
926 1 0
927 0 0
928 0 0
929 0 0
930 1 0
931 0 0
932 1 0
933 0 0
934 1 0
935 0 0
936 0 0
937 1 0
938 0 0
939 0 0
940 1 0
941 0 0
942 0 0
943 0 0
944 0 0
945 1 0
946 0 0
947 1 0
948 0 0
949 0 0
950 0 0
951 1 0
952 0 0
953 0 0
954 1 0
955 0 0
956 0 0
957 0 0
958 0 0
959 1 0
960 0 0
961 0 0
962 1 0
963 0 0
964 0 0
965 0 0
966 0 0
967 0 0
968 0 0
969 1 0
970 0 0
971 0 0
972 0 0
973 0 0
974 0 0
975 0 0
976 1 0
977 1 0
978 0 0
979 0 0
980 0 0
981 1 0
982 1 0
983 0 0
984 0 0
985 3 0
986 0 0
987 1 0
988 1 0

, , SUB.BLOCK = C

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 0
853 0 0
854 0 0
855 0 0
857 1 0
858 0 0
859 1 0
860 0 0
861 0 0
862 1 0
863 0 0
864 0 0
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 0 0
872 1 0
873 0 0
874 0 0
875 0 0
876 0 0
877 0 0
878 0 0
879 0 0
880 0 0
881 1 0
882 0 0
883 0 0
884 0 0
885 0 0
886 1 0
887 0 0
888 0 0
889 0 0
890 0 0
891 0 0
892 0 0
893 0 0
894 0 0
895 0 0
896 0 0
897 0 0
898 1 0
899 1 0
900 0 0
901 0 0
902 0 0
903 1 0
904 0 0
905 0 0
906 1 0
907 0 0
908 0 0
909 0 0
910 1 0
911 0 0
912 1 0
913 0 0
914 1 0
915 1 0
916 0 0
917 1 0
918 1 0
919 0 0
920 0 0
921 0 0
922 1 0
923 0 0
924 1 0
925 0 0
926 0 0
927 0 0
928 1 0
929 1 0
930 0 0
931 0 0
932 0 0
933 1 0
934 0 0
935 1 0
936 0 0
937 0 0
938 0 0
939 0 0
940 0 0
941 0 0
942 1 0
943 0 0
944 1 0
945 0 0
946 1 0
947 0 0
948 0 0
949 1 0
950 0 0
951 0 0
952 1 0
953 1 0
954 0 0
955 0 0
956 0 0
957 0 0
958 0 0
959 0 0
960 1 0
961 1 0
962 0 0
963 1 0
964 1 0
965 0 0
966 1 0
967 0 0
968 1 0
969 0 0
970 0 0
971 1 0
972 1 0
973 1 0
974 0 0
975 1 0
976 0 0
977 0 0
978 1 0
979 2 0
980 0 0
981 0 0
982 0 0
983 0 0
984 0 0
985 1 0
986 3 0
987 0 0
988 0 0

, , SUB.BLOCK = D

REP
ENTRY_NO 1 2
840 0 0
850 0 0
851 0 0
852 0 1
853 0 0
854 0 0
855 0 0
857 0 0
858 0 1
859 0 0
860 0 1
861 0 0
862 0 0
863 0 0
864 0 1
865 0 0
866 0 0
867 0 0
868 0 0
869 0 0
870 0 0
871 0 0
872 0 0
873 0 0
874 0 0
875 0 1
876 0 0
877 0 0
878 0 1
879 0 0
880 0 1
881 0 1
882 0 1
883 0 1
884 0 1
885 0 0
886 0 0
887 0 0
888 0 0
889 0 0
890 0 0
891 0 0
892 0 1
893 0 0
894 0 0
895 0 0
896 0 0
897 0 1
898 0 0
899 0 1
900 0 0
901 0 0
902 0 1
903 0 0
904 0 0
905 0 0
906 0 0
907 0 0
908 0 0
909 0 0
910 0 0
911 0 0
912 0 0
913 0 1
914 0 1
915 0 1
916 0 0
917 0 1
918 0 1
919 0 0
920 0 0
921 0 1
922 0 1
923 0 0
924 0 0
925 0 0
926 0 0
927 0 0
928 0 0
929 0 1
930 0 1
931 0 0
932 0 0


Can someone please give me an idea of what is going wrong??



Thanks !







r linear-regression lsmeans






share|improve this question















share|improve this question













share|improve this question




share|improve this question








edited Nov 22 '18 at 5:28







wardah m

















asked Nov 20 '18 at 0:33









wardah mwardah m

74




74













  • Can you show head(ASM_Data) in your question? It looks like ENTRY_NO may be an ID number, with only 1 observation per ENTRY_NO?

    – Marius
    Nov 20 '18 at 0:42











  • Hi~ let me add to the post.

    – wardah m
    Nov 20 '18 at 0:46











  • Where is the variable SUB.BLOCK? It’s in the model but not in the data.

    – rvl
    Nov 21 '18 at 4:47











  • Hi Sub.Plot is renamed as Sub.Block in my script.

    – wardah m
    Nov 21 '18 at 22:45



















  • Can you show head(ASM_Data) in your question? It looks like ENTRY_NO may be an ID number, with only 1 observation per ENTRY_NO?

    – Marius
    Nov 20 '18 at 0:42











  • Hi~ let me add to the post.

    – wardah m
    Nov 20 '18 at 0:46











  • Where is the variable SUB.BLOCK? It’s in the model but not in the data.

    – rvl
    Nov 21 '18 at 4:47











  • Hi Sub.Plot is renamed as Sub.Block in my script.

    – wardah m
    Nov 21 '18 at 22:45

















Can you show head(ASM_Data) in your question? It looks like ENTRY_NO may be an ID number, with only 1 observation per ENTRY_NO?

– Marius
Nov 20 '18 at 0:42





Can you show head(ASM_Data) in your question? It looks like ENTRY_NO may be an ID number, with only 1 observation per ENTRY_NO?

– Marius
Nov 20 '18 at 0:42













Hi~ let me add to the post.

– wardah m
Nov 20 '18 at 0:46





Hi~ let me add to the post.

– wardah m
Nov 20 '18 at 0:46













Where is the variable SUB.BLOCK? It’s in the model but not in the data.

– rvl
Nov 21 '18 at 4:47





Where is the variable SUB.BLOCK? It’s in the model but not in the data.

– rvl
Nov 21 '18 at 4:47













Hi Sub.Plot is renamed as Sub.Block in my script.

– wardah m
Nov 21 '18 at 22:45





Hi Sub.Plot is renamed as Sub.Block in my script.

– wardah m
Nov 21 '18 at 22:45












2 Answers
2






active

oldest

votes


















0














I've been able to create a situation like this. Consider this dataset:



> junk
trt rep blk y
1 A 1 1 -1.17415687
2 B 1 1 -0.20084854
3 C 1 1 0.64797806
4 A 1 2 -1.69371434
5 B 1 2 -0.35835442
6 C 1 2 1.35718782
7 A 1 3 0.20510482
8 B 1 3 1.00857651
9 C 1 3 -0.20553167
10 A 2 4 0.31261523
11 B 2 4 0.47989115
12 C 2 4 1.27574085
13 A 2 5 -0.79209520
14 B 2 5 1.07151315
15 C 2 5 -0.04222769
16 A 2 6 -0.80571767
17 B 2 6 0.80442988
18 C 2 6 1.73526561


This has 6 complete blocks, separately labeled with 3 blocks per rep. Not obvious, but true, is that rep is a numeric variable having values 1 and 2, while blk is a factor having 6 levels 1 -- 6:



> sapply(junk, class)
trt rep blk y
"factor" "numeric" "factor" "numeric"


With this complete dataset, I have no problem obtaining EMMs for modeling situations parallel to what was used in the original posting. However, if I use only a subset of these data, it is different. Consider:



> samp
[1] 1 2 3 5 8 11 13 15 16

> junk.lm = lm(y ~ trt + rep + blk, data = junk, subset = samp)
> emmeans(junk.lm, "trt")
trt emmean SE df asymp.LCL asymp.UCL
A nonEst NA NA NA NA
B nonEst NA NA NA NA
C nonEst NA NA NA NA

Results are averaged over the levels of: blk
Confidence level used: 0.95


Again, recall that rep is numeric in this model. If instead, I make rep a factor:



> junk.lmf = lm(y ~ trt + factor(rep) + blk, data = junk, subset = samp)
> emmeans(junk.lmf, "trt")
NOTE: A nesting structure was detected in the fitted model:
blk %in% rep
If this is incorrect, re-run or update with `nesting` specified
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk, rep
Confidence level used: 0.95


We get non-NA estimates, in part because it is able to detect the fact that blk is nested in rep, and thus performs the EMM computations separately in each rep. Note in the annotations in this last output that averaging is done over the 2 reps and 6 blocks; whereas in fiber.lm averaging is done only over blocks, while rep, a numeric variable, is set at its average. Compare the reference grids for the two models:



> ref_grid(junk.lm)
'emmGrid' object with variables:
trt = A, B, C
rep = 1.4444
blk = 1, 2, 3, 4, 5, 6

> ref_grid(junk.lmf)
'emmGrid' object with variables:
trt = A, B, C
rep = 1, 2
blk = 1, 2, 3, 4, 5, 6
Nesting structure: blk %in% rep


An additional option is to avoid the nesting issue by simply omitting rep from the model:



> junk.lm.norep = lm(y ~ trt + blk, data = junk, subset = samp)
> emmeans(junk.lm.norep, "trt")
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk
Confidence level used: 0.95


Note that exactly the same results are produced. The reason is the levels of blk already predict the levels of rep, so there is no need for it to be in the model.



In summary:




  • The situation is due in part to the fact that there are missing data

  • and in part because rep was in the model as a numeric predictor rather than a factor.

  • In your situation, I suggest re-fitting the model with factor(REP) instead of REPas a numeric predictor. This may be enough to produce estimates.

  • If, indeed, as in my example, the SUB.BLOCK levels predict the REP levels, just leave REP out of the model altogether.






share|improve this answer
























  • Great ! This helped alot! I will keep my REP but change it to a factor. Appreciate you patience and guidance!

    – wardah m
    Nov 22 '18 at 21:07



















0














EMMs are obtained by averaging predictions over 2 reps and 5 blocks (or maybe more?). Look at



coef(ASM_YIELD_1)


If any of the rep or block effects are NA, then you can’t estimate all of the rep or block effects, and that makes the average of them non-estimable.



You can see exactly which factor combinations are non-estimable by doing:



summary(ref_grid(ASM_YIELD_1))


addendum



Here is a reformatting of the tables I requested in comments:



ENTRY   ---------- BLOCK -------------
NO A B C D

840 0 0 0 0 0 0 0 0
850 0 0 0 0 0 0 0 0
851 0 0 0 0 0 0 0 0
852 0 0 0 0 0 0 0 1
853 0 0 1 0 0 0 0 0
854 0 0 0 0 0 0 0 0
855 0 0 0 0 0 0 0 0
857 0 0 0 0 1 0 0 0
858 0 0 0 0 0 0 0 1
859 0 0 0 0 1 0 0 0
... etc ...


This is extremely sparse data. I think there are two more blocks not shown. But I see very few instances where a given ENTRY_NO is observed in more than one rep or block. So I think it is seriously over-fitting to try to account for rep or block effects in this model.



MAYBE omitting REP from the model will make it work. MAYBE re-fitting the model with factor(REP) in place of REP will enable emmeans to detect a nesting structure. Otherwise, there's some really subtle dependence in the blocking structure and treatments, and I don't know what to suggest.






share|improve this answer


























  • Hi ~ I cannot tell much from this but I get the same result for my other traits as well (flowering, height etc.) each emmean comparison gets a nonEst. I just think this is really strange.

    – wardah m
    Nov 22 '18 at 1:51











  • Yes — changing the response variable will make no difference whatsoever, because this is due to linear dependence in the factor levels. Did you print that summary of the reference grid that I suggested?

    – rvl
    Nov 22 '18 at 2:12











  • @rvi coef(ASM_YIELD_1) (Intercept) ENTRY_NO982 ENTRY_NO983 ENTRY_NO984 ENTRY_NO985 1.39184456 -20.85391334 16.60855561 7.69139728 9.63711948 -2.29160352 -18.46223843 -26.07993162 ENTRY_NO986 ENTRY_NO987 ENTRY_NO988 REP SUB.BLOCKB SUB.BLOCKC SUB.BLOCKD SUB.BLOCKE 8.97955019 6.69982467 -47.13796781 37.71682387 13.06799679 5.98303157 6.10034916 3.78288879 SUB.BLOCKF NA

    – wardah m
    Nov 22 '18 at 3:21













  • @rvi Yes I seem to have it for Block F , how can I change the code to omit the na so that this can work?

    – wardah m
    Nov 22 '18 at 3:24











  • I’d suggest taking REP out of the model. My guess it is providing completely redundant info and that’s why things can’t be estimated. But I am just guessing because you still haven’t provided the second set of details I mentioned.

    – rvl
    Nov 22 '18 at 4:06











Your Answer






StackExchange.ifUsing("editor", function () {
StackExchange.using("externalEditor", function () {
StackExchange.using("snippets", function () {
StackExchange.snippets.init();
});
});
}, "code-snippets");

StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "1"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);

StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});

function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});


}
});














draft saved

draft discarded


















StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53384585%2fr-ls-means-analysis-produces-nas%23new-answer', 'question_page');
}
);

Post as a guest















Required, but never shown

























2 Answers
2






active

oldest

votes








2 Answers
2






active

oldest

votes









active

oldest

votes






active

oldest

votes









0














I've been able to create a situation like this. Consider this dataset:



> junk
trt rep blk y
1 A 1 1 -1.17415687
2 B 1 1 -0.20084854
3 C 1 1 0.64797806
4 A 1 2 -1.69371434
5 B 1 2 -0.35835442
6 C 1 2 1.35718782
7 A 1 3 0.20510482
8 B 1 3 1.00857651
9 C 1 3 -0.20553167
10 A 2 4 0.31261523
11 B 2 4 0.47989115
12 C 2 4 1.27574085
13 A 2 5 -0.79209520
14 B 2 5 1.07151315
15 C 2 5 -0.04222769
16 A 2 6 -0.80571767
17 B 2 6 0.80442988
18 C 2 6 1.73526561


This has 6 complete blocks, separately labeled with 3 blocks per rep. Not obvious, but true, is that rep is a numeric variable having values 1 and 2, while blk is a factor having 6 levels 1 -- 6:



> sapply(junk, class)
trt rep blk y
"factor" "numeric" "factor" "numeric"


With this complete dataset, I have no problem obtaining EMMs for modeling situations parallel to what was used in the original posting. However, if I use only a subset of these data, it is different. Consider:



> samp
[1] 1 2 3 5 8 11 13 15 16

> junk.lm = lm(y ~ trt + rep + blk, data = junk, subset = samp)
> emmeans(junk.lm, "trt")
trt emmean SE df asymp.LCL asymp.UCL
A nonEst NA NA NA NA
B nonEst NA NA NA NA
C nonEst NA NA NA NA

Results are averaged over the levels of: blk
Confidence level used: 0.95


Again, recall that rep is numeric in this model. If instead, I make rep a factor:



> junk.lmf = lm(y ~ trt + factor(rep) + blk, data = junk, subset = samp)
> emmeans(junk.lmf, "trt")
NOTE: A nesting structure was detected in the fitted model:
blk %in% rep
If this is incorrect, re-run or update with `nesting` specified
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk, rep
Confidence level used: 0.95


We get non-NA estimates, in part because it is able to detect the fact that blk is nested in rep, and thus performs the EMM computations separately in each rep. Note in the annotations in this last output that averaging is done over the 2 reps and 6 blocks; whereas in fiber.lm averaging is done only over blocks, while rep, a numeric variable, is set at its average. Compare the reference grids for the two models:



> ref_grid(junk.lm)
'emmGrid' object with variables:
trt = A, B, C
rep = 1.4444
blk = 1, 2, 3, 4, 5, 6

> ref_grid(junk.lmf)
'emmGrid' object with variables:
trt = A, B, C
rep = 1, 2
blk = 1, 2, 3, 4, 5, 6
Nesting structure: blk %in% rep


An additional option is to avoid the nesting issue by simply omitting rep from the model:



> junk.lm.norep = lm(y ~ trt + blk, data = junk, subset = samp)
> emmeans(junk.lm.norep, "trt")
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk
Confidence level used: 0.95


Note that exactly the same results are produced. The reason is the levels of blk already predict the levels of rep, so there is no need for it to be in the model.



In summary:




  • The situation is due in part to the fact that there are missing data

  • and in part because rep was in the model as a numeric predictor rather than a factor.

  • In your situation, I suggest re-fitting the model with factor(REP) instead of REPas a numeric predictor. This may be enough to produce estimates.

  • If, indeed, as in my example, the SUB.BLOCK levels predict the REP levels, just leave REP out of the model altogether.






share|improve this answer
























  • Great ! This helped alot! I will keep my REP but change it to a factor. Appreciate you patience and guidance!

    – wardah m
    Nov 22 '18 at 21:07
















0














I've been able to create a situation like this. Consider this dataset:



> junk
trt rep blk y
1 A 1 1 -1.17415687
2 B 1 1 -0.20084854
3 C 1 1 0.64797806
4 A 1 2 -1.69371434
5 B 1 2 -0.35835442
6 C 1 2 1.35718782
7 A 1 3 0.20510482
8 B 1 3 1.00857651
9 C 1 3 -0.20553167
10 A 2 4 0.31261523
11 B 2 4 0.47989115
12 C 2 4 1.27574085
13 A 2 5 -0.79209520
14 B 2 5 1.07151315
15 C 2 5 -0.04222769
16 A 2 6 -0.80571767
17 B 2 6 0.80442988
18 C 2 6 1.73526561


This has 6 complete blocks, separately labeled with 3 blocks per rep. Not obvious, but true, is that rep is a numeric variable having values 1 and 2, while blk is a factor having 6 levels 1 -- 6:



> sapply(junk, class)
trt rep blk y
"factor" "numeric" "factor" "numeric"


With this complete dataset, I have no problem obtaining EMMs for modeling situations parallel to what was used in the original posting. However, if I use only a subset of these data, it is different. Consider:



> samp
[1] 1 2 3 5 8 11 13 15 16

> junk.lm = lm(y ~ trt + rep + blk, data = junk, subset = samp)
> emmeans(junk.lm, "trt")
trt emmean SE df asymp.LCL asymp.UCL
A nonEst NA NA NA NA
B nonEst NA NA NA NA
C nonEst NA NA NA NA

Results are averaged over the levels of: blk
Confidence level used: 0.95


Again, recall that rep is numeric in this model. If instead, I make rep a factor:



> junk.lmf = lm(y ~ trt + factor(rep) + blk, data = junk, subset = samp)
> emmeans(junk.lmf, "trt")
NOTE: A nesting structure was detected in the fitted model:
blk %in% rep
If this is incorrect, re-run or update with `nesting` specified
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk, rep
Confidence level used: 0.95


We get non-NA estimates, in part because it is able to detect the fact that blk is nested in rep, and thus performs the EMM computations separately in each rep. Note in the annotations in this last output that averaging is done over the 2 reps and 6 blocks; whereas in fiber.lm averaging is done only over blocks, while rep, a numeric variable, is set at its average. Compare the reference grids for the two models:



> ref_grid(junk.lm)
'emmGrid' object with variables:
trt = A, B, C
rep = 1.4444
blk = 1, 2, 3, 4, 5, 6

> ref_grid(junk.lmf)
'emmGrid' object with variables:
trt = A, B, C
rep = 1, 2
blk = 1, 2, 3, 4, 5, 6
Nesting structure: blk %in% rep


An additional option is to avoid the nesting issue by simply omitting rep from the model:



> junk.lm.norep = lm(y ~ trt + blk, data = junk, subset = samp)
> emmeans(junk.lm.norep, "trt")
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk
Confidence level used: 0.95


Note that exactly the same results are produced. The reason is the levels of blk already predict the levels of rep, so there is no need for it to be in the model.



In summary:




  • The situation is due in part to the fact that there are missing data

  • and in part because rep was in the model as a numeric predictor rather than a factor.

  • In your situation, I suggest re-fitting the model with factor(REP) instead of REPas a numeric predictor. This may be enough to produce estimates.

  • If, indeed, as in my example, the SUB.BLOCK levels predict the REP levels, just leave REP out of the model altogether.






share|improve this answer
























  • Great ! This helped alot! I will keep my REP but change it to a factor. Appreciate you patience and guidance!

    – wardah m
    Nov 22 '18 at 21:07














0












0








0







I've been able to create a situation like this. Consider this dataset:



> junk
trt rep blk y
1 A 1 1 -1.17415687
2 B 1 1 -0.20084854
3 C 1 1 0.64797806
4 A 1 2 -1.69371434
5 B 1 2 -0.35835442
6 C 1 2 1.35718782
7 A 1 3 0.20510482
8 B 1 3 1.00857651
9 C 1 3 -0.20553167
10 A 2 4 0.31261523
11 B 2 4 0.47989115
12 C 2 4 1.27574085
13 A 2 5 -0.79209520
14 B 2 5 1.07151315
15 C 2 5 -0.04222769
16 A 2 6 -0.80571767
17 B 2 6 0.80442988
18 C 2 6 1.73526561


This has 6 complete blocks, separately labeled with 3 blocks per rep. Not obvious, but true, is that rep is a numeric variable having values 1 and 2, while blk is a factor having 6 levels 1 -- 6:



> sapply(junk, class)
trt rep blk y
"factor" "numeric" "factor" "numeric"


With this complete dataset, I have no problem obtaining EMMs for modeling situations parallel to what was used in the original posting. However, if I use only a subset of these data, it is different. Consider:



> samp
[1] 1 2 3 5 8 11 13 15 16

> junk.lm = lm(y ~ trt + rep + blk, data = junk, subset = samp)
> emmeans(junk.lm, "trt")
trt emmean SE df asymp.LCL asymp.UCL
A nonEst NA NA NA NA
B nonEst NA NA NA NA
C nonEst NA NA NA NA

Results are averaged over the levels of: blk
Confidence level used: 0.95


Again, recall that rep is numeric in this model. If instead, I make rep a factor:



> junk.lmf = lm(y ~ trt + factor(rep) + blk, data = junk, subset = samp)
> emmeans(junk.lmf, "trt")
NOTE: A nesting structure was detected in the fitted model:
blk %in% rep
If this is incorrect, re-run or update with `nesting` specified
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk, rep
Confidence level used: 0.95


We get non-NA estimates, in part because it is able to detect the fact that blk is nested in rep, and thus performs the EMM computations separately in each rep. Note in the annotations in this last output that averaging is done over the 2 reps and 6 blocks; whereas in fiber.lm averaging is done only over blocks, while rep, a numeric variable, is set at its average. Compare the reference grids for the two models:



> ref_grid(junk.lm)
'emmGrid' object with variables:
trt = A, B, C
rep = 1.4444
blk = 1, 2, 3, 4, 5, 6

> ref_grid(junk.lmf)
'emmGrid' object with variables:
trt = A, B, C
rep = 1, 2
blk = 1, 2, 3, 4, 5, 6
Nesting structure: blk %in% rep


An additional option is to avoid the nesting issue by simply omitting rep from the model:



> junk.lm.norep = lm(y ~ trt + blk, data = junk, subset = samp)
> emmeans(junk.lm.norep, "trt")
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk
Confidence level used: 0.95


Note that exactly the same results are produced. The reason is the levels of blk already predict the levels of rep, so there is no need for it to be in the model.



In summary:




  • The situation is due in part to the fact that there are missing data

  • and in part because rep was in the model as a numeric predictor rather than a factor.

  • In your situation, I suggest re-fitting the model with factor(REP) instead of REPas a numeric predictor. This may be enough to produce estimates.

  • If, indeed, as in my example, the SUB.BLOCK levels predict the REP levels, just leave REP out of the model altogether.






share|improve this answer













I've been able to create a situation like this. Consider this dataset:



> junk
trt rep blk y
1 A 1 1 -1.17415687
2 B 1 1 -0.20084854
3 C 1 1 0.64797806
4 A 1 2 -1.69371434
5 B 1 2 -0.35835442
6 C 1 2 1.35718782
7 A 1 3 0.20510482
8 B 1 3 1.00857651
9 C 1 3 -0.20553167
10 A 2 4 0.31261523
11 B 2 4 0.47989115
12 C 2 4 1.27574085
13 A 2 5 -0.79209520
14 B 2 5 1.07151315
15 C 2 5 -0.04222769
16 A 2 6 -0.80571767
17 B 2 6 0.80442988
18 C 2 6 1.73526561


This has 6 complete blocks, separately labeled with 3 blocks per rep. Not obvious, but true, is that rep is a numeric variable having values 1 and 2, while blk is a factor having 6 levels 1 -- 6:



> sapply(junk, class)
trt rep blk y
"factor" "numeric" "factor" "numeric"


With this complete dataset, I have no problem obtaining EMMs for modeling situations parallel to what was used in the original posting. However, if I use only a subset of these data, it is different. Consider:



> samp
[1] 1 2 3 5 8 11 13 15 16

> junk.lm = lm(y ~ trt + rep + blk, data = junk, subset = samp)
> emmeans(junk.lm, "trt")
trt emmean SE df asymp.LCL asymp.UCL
A nonEst NA NA NA NA
B nonEst NA NA NA NA
C nonEst NA NA NA NA

Results are averaged over the levels of: blk
Confidence level used: 0.95


Again, recall that rep is numeric in this model. If instead, I make rep a factor:



> junk.lmf = lm(y ~ trt + factor(rep) + blk, data = junk, subset = samp)
> emmeans(junk.lmf, "trt")
NOTE: A nesting structure was detected in the fitted model:
blk %in% rep
If this is incorrect, re-run or update with `nesting` specified
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk, rep
Confidence level used: 0.95


We get non-NA estimates, in part because it is able to detect the fact that blk is nested in rep, and thus performs the EMM computations separately in each rep. Note in the annotations in this last output that averaging is done over the 2 reps and 6 blocks; whereas in fiber.lm averaging is done only over blocks, while rep, a numeric variable, is set at its average. Compare the reference grids for the two models:



> ref_grid(junk.lm)
'emmGrid' object with variables:
trt = A, B, C
rep = 1.4444
blk = 1, 2, 3, 4, 5, 6

> ref_grid(junk.lmf)
'emmGrid' object with variables:
trt = A, B, C
rep = 1, 2
blk = 1, 2, 3, 4, 5, 6
Nesting structure: blk %in% rep


An additional option is to avoid the nesting issue by simply omitting rep from the model:



> junk.lm.norep = lm(y ~ trt + blk, data = junk, subset = samp)
> emmeans(junk.lm.norep, "trt")
trt emmean SE df lower.CL upper.CL
A -0.6262635 0.4707099 1 -6.607200 5.354673
B 0.0789780 0.3546191 1 -4.426885 4.584841
C 0.6597377 0.5191092 1 -5.936170 7.255646

Results are averaged over the levels of: blk
Confidence level used: 0.95


Note that exactly the same results are produced. The reason is the levels of blk already predict the levels of rep, so there is no need for it to be in the model.



In summary:




  • The situation is due in part to the fact that there are missing data

  • and in part because rep was in the model as a numeric predictor rather than a factor.

  • In your situation, I suggest re-fitting the model with factor(REP) instead of REPas a numeric predictor. This may be enough to produce estimates.

  • If, indeed, as in my example, the SUB.BLOCK levels predict the REP levels, just leave REP out of the model altogether.







share|improve this answer












share|improve this answer



share|improve this answer










answered Nov 22 '18 at 17:41









rvlrvl

1,9552613




1,9552613













  • Great ! This helped alot! I will keep my REP but change it to a factor. Appreciate you patience and guidance!

    – wardah m
    Nov 22 '18 at 21:07



















  • Great ! This helped alot! I will keep my REP but change it to a factor. Appreciate you patience and guidance!

    – wardah m
    Nov 22 '18 at 21:07

















Great ! This helped alot! I will keep my REP but change it to a factor. Appreciate you patience and guidance!

– wardah m
Nov 22 '18 at 21:07





Great ! This helped alot! I will keep my REP but change it to a factor. Appreciate you patience and guidance!

– wardah m
Nov 22 '18 at 21:07













0














EMMs are obtained by averaging predictions over 2 reps and 5 blocks (or maybe more?). Look at



coef(ASM_YIELD_1)


If any of the rep or block effects are NA, then you can’t estimate all of the rep or block effects, and that makes the average of them non-estimable.



You can see exactly which factor combinations are non-estimable by doing:



summary(ref_grid(ASM_YIELD_1))


addendum



Here is a reformatting of the tables I requested in comments:



ENTRY   ---------- BLOCK -------------
NO A B C D

840 0 0 0 0 0 0 0 0
850 0 0 0 0 0 0 0 0
851 0 0 0 0 0 0 0 0
852 0 0 0 0 0 0 0 1
853 0 0 1 0 0 0 0 0
854 0 0 0 0 0 0 0 0
855 0 0 0 0 0 0 0 0
857 0 0 0 0 1 0 0 0
858 0 0 0 0 0 0 0 1
859 0 0 0 0 1 0 0 0
... etc ...


This is extremely sparse data. I think there are two more blocks not shown. But I see very few instances where a given ENTRY_NO is observed in more than one rep or block. So I think it is seriously over-fitting to try to account for rep or block effects in this model.



MAYBE omitting REP from the model will make it work. MAYBE re-fitting the model with factor(REP) in place of REP will enable emmeans to detect a nesting structure. Otherwise, there's some really subtle dependence in the blocking structure and treatments, and I don't know what to suggest.






share|improve this answer


























  • Hi ~ I cannot tell much from this but I get the same result for my other traits as well (flowering, height etc.) each emmean comparison gets a nonEst. I just think this is really strange.

    – wardah m
    Nov 22 '18 at 1:51











  • Yes — changing the response variable will make no difference whatsoever, because this is due to linear dependence in the factor levels. Did you print that summary of the reference grid that I suggested?

    – rvl
    Nov 22 '18 at 2:12











  • @rvi coef(ASM_YIELD_1) (Intercept) ENTRY_NO982 ENTRY_NO983 ENTRY_NO984 ENTRY_NO985 1.39184456 -20.85391334 16.60855561 7.69139728 9.63711948 -2.29160352 -18.46223843 -26.07993162 ENTRY_NO986 ENTRY_NO987 ENTRY_NO988 REP SUB.BLOCKB SUB.BLOCKC SUB.BLOCKD SUB.BLOCKE 8.97955019 6.69982467 -47.13796781 37.71682387 13.06799679 5.98303157 6.10034916 3.78288879 SUB.BLOCKF NA

    – wardah m
    Nov 22 '18 at 3:21













  • @rvi Yes I seem to have it for Block F , how can I change the code to omit the na so that this can work?

    – wardah m
    Nov 22 '18 at 3:24











  • I’d suggest taking REP out of the model. My guess it is providing completely redundant info and that’s why things can’t be estimated. But I am just guessing because you still haven’t provided the second set of details I mentioned.

    – rvl
    Nov 22 '18 at 4:06
















0














EMMs are obtained by averaging predictions over 2 reps and 5 blocks (or maybe more?). Look at



coef(ASM_YIELD_1)


If any of the rep or block effects are NA, then you can’t estimate all of the rep or block effects, and that makes the average of them non-estimable.



You can see exactly which factor combinations are non-estimable by doing:



summary(ref_grid(ASM_YIELD_1))


addendum



Here is a reformatting of the tables I requested in comments:



ENTRY   ---------- BLOCK -------------
NO A B C D

840 0 0 0 0 0 0 0 0
850 0 0 0 0 0 0 0 0
851 0 0 0 0 0 0 0 0
852 0 0 0 0 0 0 0 1
853 0 0 1 0 0 0 0 0
854 0 0 0 0 0 0 0 0
855 0 0 0 0 0 0 0 0
857 0 0 0 0 1 0 0 0
858 0 0 0 0 0 0 0 1
859 0 0 0 0 1 0 0 0
... etc ...


This is extremely sparse data. I think there are two more blocks not shown. But I see very few instances where a given ENTRY_NO is observed in more than one rep or block. So I think it is seriously over-fitting to try to account for rep or block effects in this model.



MAYBE omitting REP from the model will make it work. MAYBE re-fitting the model with factor(REP) in place of REP will enable emmeans to detect a nesting structure. Otherwise, there's some really subtle dependence in the blocking structure and treatments, and I don't know what to suggest.






share|improve this answer


























  • Hi ~ I cannot tell much from this but I get the same result for my other traits as well (flowering, height etc.) each emmean comparison gets a nonEst. I just think this is really strange.

    – wardah m
    Nov 22 '18 at 1:51











  • Yes — changing the response variable will make no difference whatsoever, because this is due to linear dependence in the factor levels. Did you print that summary of the reference grid that I suggested?

    – rvl
    Nov 22 '18 at 2:12











  • @rvi coef(ASM_YIELD_1) (Intercept) ENTRY_NO982 ENTRY_NO983 ENTRY_NO984 ENTRY_NO985 1.39184456 -20.85391334 16.60855561 7.69139728 9.63711948 -2.29160352 -18.46223843 -26.07993162 ENTRY_NO986 ENTRY_NO987 ENTRY_NO988 REP SUB.BLOCKB SUB.BLOCKC SUB.BLOCKD SUB.BLOCKE 8.97955019 6.69982467 -47.13796781 37.71682387 13.06799679 5.98303157 6.10034916 3.78288879 SUB.BLOCKF NA

    – wardah m
    Nov 22 '18 at 3:21













  • @rvi Yes I seem to have it for Block F , how can I change the code to omit the na so that this can work?

    – wardah m
    Nov 22 '18 at 3:24











  • I’d suggest taking REP out of the model. My guess it is providing completely redundant info and that’s why things can’t be estimated. But I am just guessing because you still haven’t provided the second set of details I mentioned.

    – rvl
    Nov 22 '18 at 4:06














0












0








0







EMMs are obtained by averaging predictions over 2 reps and 5 blocks (or maybe more?). Look at



coef(ASM_YIELD_1)


If any of the rep or block effects are NA, then you can’t estimate all of the rep or block effects, and that makes the average of them non-estimable.



You can see exactly which factor combinations are non-estimable by doing:



summary(ref_grid(ASM_YIELD_1))


addendum



Here is a reformatting of the tables I requested in comments:



ENTRY   ---------- BLOCK -------------
NO A B C D

840 0 0 0 0 0 0 0 0
850 0 0 0 0 0 0 0 0
851 0 0 0 0 0 0 0 0
852 0 0 0 0 0 0 0 1
853 0 0 1 0 0 0 0 0
854 0 0 0 0 0 0 0 0
855 0 0 0 0 0 0 0 0
857 0 0 0 0 1 0 0 0
858 0 0 0 0 0 0 0 1
859 0 0 0 0 1 0 0 0
... etc ...


This is extremely sparse data. I think there are two more blocks not shown. But I see very few instances where a given ENTRY_NO is observed in more than one rep or block. So I think it is seriously over-fitting to try to account for rep or block effects in this model.



MAYBE omitting REP from the model will make it work. MAYBE re-fitting the model with factor(REP) in place of REP will enable emmeans to detect a nesting structure. Otherwise, there's some really subtle dependence in the blocking structure and treatments, and I don't know what to suggest.






share|improve this answer















EMMs are obtained by averaging predictions over 2 reps and 5 blocks (or maybe more?). Look at



coef(ASM_YIELD_1)


If any of the rep or block effects are NA, then you can’t estimate all of the rep or block effects, and that makes the average of them non-estimable.



You can see exactly which factor combinations are non-estimable by doing:



summary(ref_grid(ASM_YIELD_1))


addendum



Here is a reformatting of the tables I requested in comments:



ENTRY   ---------- BLOCK -------------
NO A B C D

840 0 0 0 0 0 0 0 0
850 0 0 0 0 0 0 0 0
851 0 0 0 0 0 0 0 0
852 0 0 0 0 0 0 0 1
853 0 0 1 0 0 0 0 0
854 0 0 0 0 0 0 0 0
855 0 0 0 0 0 0 0 0
857 0 0 0 0 1 0 0 0
858 0 0 0 0 0 0 0 1
859 0 0 0 0 1 0 0 0
... etc ...


This is extremely sparse data. I think there are two more blocks not shown. But I see very few instances where a given ENTRY_NO is observed in more than one rep or block. So I think it is seriously over-fitting to try to account for rep or block effects in this model.



MAYBE omitting REP from the model will make it work. MAYBE re-fitting the model with factor(REP) in place of REP will enable emmeans to detect a nesting structure. Otherwise, there's some really subtle dependence in the blocking structure and treatments, and I don't know what to suggest.







share|improve this answer














share|improve this answer



share|improve this answer








edited Nov 22 '18 at 6:21

























answered Nov 21 '18 at 4:57









rvlrvl

1,9552613




1,9552613













  • Hi ~ I cannot tell much from this but I get the same result for my other traits as well (flowering, height etc.) each emmean comparison gets a nonEst. I just think this is really strange.

    – wardah m
    Nov 22 '18 at 1:51











  • Yes — changing the response variable will make no difference whatsoever, because this is due to linear dependence in the factor levels. Did you print that summary of the reference grid that I suggested?

    – rvl
    Nov 22 '18 at 2:12











  • @rvi coef(ASM_YIELD_1) (Intercept) ENTRY_NO982 ENTRY_NO983 ENTRY_NO984 ENTRY_NO985 1.39184456 -20.85391334 16.60855561 7.69139728 9.63711948 -2.29160352 -18.46223843 -26.07993162 ENTRY_NO986 ENTRY_NO987 ENTRY_NO988 REP SUB.BLOCKB SUB.BLOCKC SUB.BLOCKD SUB.BLOCKE 8.97955019 6.69982467 -47.13796781 37.71682387 13.06799679 5.98303157 6.10034916 3.78288879 SUB.BLOCKF NA

    – wardah m
    Nov 22 '18 at 3:21













  • @rvi Yes I seem to have it for Block F , how can I change the code to omit the na so that this can work?

    – wardah m
    Nov 22 '18 at 3:24











  • I’d suggest taking REP out of the model. My guess it is providing completely redundant info and that’s why things can’t be estimated. But I am just guessing because you still haven’t provided the second set of details I mentioned.

    – rvl
    Nov 22 '18 at 4:06



















  • Hi ~ I cannot tell much from this but I get the same result for my other traits as well (flowering, height etc.) each emmean comparison gets a nonEst. I just think this is really strange.

    – wardah m
    Nov 22 '18 at 1:51











  • Yes — changing the response variable will make no difference whatsoever, because this is due to linear dependence in the factor levels. Did you print that summary of the reference grid that I suggested?

    – rvl
    Nov 22 '18 at 2:12











  • @rvi coef(ASM_YIELD_1) (Intercept) ENTRY_NO982 ENTRY_NO983 ENTRY_NO984 ENTRY_NO985 1.39184456 -20.85391334 16.60855561 7.69139728 9.63711948 -2.29160352 -18.46223843 -26.07993162 ENTRY_NO986 ENTRY_NO987 ENTRY_NO988 REP SUB.BLOCKB SUB.BLOCKC SUB.BLOCKD SUB.BLOCKE 8.97955019 6.69982467 -47.13796781 37.71682387 13.06799679 5.98303157 6.10034916 3.78288879 SUB.BLOCKF NA

    – wardah m
    Nov 22 '18 at 3:21













  • @rvi Yes I seem to have it for Block F , how can I change the code to omit the na so that this can work?

    – wardah m
    Nov 22 '18 at 3:24











  • I’d suggest taking REP out of the model. My guess it is providing completely redundant info and that’s why things can’t be estimated. But I am just guessing because you still haven’t provided the second set of details I mentioned.

    – rvl
    Nov 22 '18 at 4:06

















Hi ~ I cannot tell much from this but I get the same result for my other traits as well (flowering, height etc.) each emmean comparison gets a nonEst. I just think this is really strange.

– wardah m
Nov 22 '18 at 1:51





Hi ~ I cannot tell much from this but I get the same result for my other traits as well (flowering, height etc.) each emmean comparison gets a nonEst. I just think this is really strange.

– wardah m
Nov 22 '18 at 1:51













Yes — changing the response variable will make no difference whatsoever, because this is due to linear dependence in the factor levels. Did you print that summary of the reference grid that I suggested?

– rvl
Nov 22 '18 at 2:12





Yes — changing the response variable will make no difference whatsoever, because this is due to linear dependence in the factor levels. Did you print that summary of the reference grid that I suggested?

– rvl
Nov 22 '18 at 2:12













@rvi coef(ASM_YIELD_1) (Intercept) ENTRY_NO982 ENTRY_NO983 ENTRY_NO984 ENTRY_NO985 1.39184456 -20.85391334 16.60855561 7.69139728 9.63711948 -2.29160352 -18.46223843 -26.07993162 ENTRY_NO986 ENTRY_NO987 ENTRY_NO988 REP SUB.BLOCKB SUB.BLOCKC SUB.BLOCKD SUB.BLOCKE 8.97955019 6.69982467 -47.13796781 37.71682387 13.06799679 5.98303157 6.10034916 3.78288879 SUB.BLOCKF NA

– wardah m
Nov 22 '18 at 3:21







@rvi coef(ASM_YIELD_1) (Intercept) ENTRY_NO982 ENTRY_NO983 ENTRY_NO984 ENTRY_NO985 1.39184456 -20.85391334 16.60855561 7.69139728 9.63711948 -2.29160352 -18.46223843 -26.07993162 ENTRY_NO986 ENTRY_NO987 ENTRY_NO988 REP SUB.BLOCKB SUB.BLOCKC SUB.BLOCKD SUB.BLOCKE 8.97955019 6.69982467 -47.13796781 37.71682387 13.06799679 5.98303157 6.10034916 3.78288879 SUB.BLOCKF NA

– wardah m
Nov 22 '18 at 3:21















@rvi Yes I seem to have it for Block F , how can I change the code to omit the na so that this can work?

– wardah m
Nov 22 '18 at 3:24





@rvi Yes I seem to have it for Block F , how can I change the code to omit the na so that this can work?

– wardah m
Nov 22 '18 at 3:24













I’d suggest taking REP out of the model. My guess it is providing completely redundant info and that’s why things can’t be estimated. But I am just guessing because you still haven’t provided the second set of details I mentioned.

– rvl
Nov 22 '18 at 4:06





I’d suggest taking REP out of the model. My guess it is providing completely redundant info and that’s why things can’t be estimated. But I am just guessing because you still haven’t provided the second set of details I mentioned.

– rvl
Nov 22 '18 at 4:06


















draft saved

draft discarded




















































Thanks for contributing an answer to Stack Overflow!


  • Please be sure to answer the question. Provide details and share your research!

But avoid



  • Asking for help, clarification, or responding to other answers.

  • Making statements based on opinion; back them up with references or personal experience.


To learn more, see our tips on writing great answers.




draft saved


draft discarded














StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53384585%2fr-ls-means-analysis-produces-nas%23new-answer', 'question_page');
}
);

Post as a guest















Required, but never shown





















































Required, but never shown














Required, but never shown












Required, but never shown







Required, but never shown

































Required, but never shown














Required, but never shown












Required, but never shown







Required, but never shown







Popular posts from this blog

Guess what letter conforming each word

Port of Spain

Run scheduled task as local user group (not BUILTIN)