Murray Jorgensen wrote:
Hi Peter,
there is no problem if the missing cell is not in the first row or
column: the corresponding interaction parameter is omitted. In my case
the data in the (1,4) cell is missing. What results is clear to me now:
the (3,4) interaction parameter is dropped so
Hi Peter,
there is no problem if the missing cell is not in the first row or
column: the corresponding interaction parameter is omitted. In my case
the data in the (1,4) cell is missing. What results is clear to me now:
the (3,4) interaction parameter is dropped so that (Intercept) + Biv
now
I am wondering how to interpret the parameter estimates that lm()
reports in this sort of situation:
y = round(rnorm(n=24,mean=5,sd=2),2)
A = gl(3,2,24,labels=c(one,two,three))
B = gl(4,6,24,labels=c(i,ii,iii,iv))
# Make both observations for A=1, B=4 missing
y[19] = NA
y[20] = NA
Does this help at all?
after your code...
contrasts(A)
two three
one 0 0
two 1 0
three 0 1
contrasts(B)
ii iii iv
i0 0 0
ii 1 0 0
iii 0 1 0
iv 0 0 1
contrasts(A:B)
one:ii one:iii one:iv two:i two:ii two:iii two:iv three:i
I am wondering how to interpret the parameter estimates that lm()
reports in this sort of situation:
y = round(rnorm(n=24,mean=5,sd=2),2)
A = gl(3,2,24,labels=c(one,two,three))
B = gl(4,6,24,labels=c(i,ii,iii,iv))
# Make both observations for A=1, B=4 missing
y[19] = NA
y[20] = NA
Murray Jorgensen wrote:
I am wondering how to interpret the parameter estimates that lm()
reports in this sort of situation:
y = round(rnorm(n=24,mean=5,sd=2),2)
A = gl(3,2,24,labels=c(one,two,three))
B = gl(4,6,24,labels=c(i,ii,iii,iv))
# Make both observations for A=1, B=4 missing
y[19] = NA
6 matches
Mail list logo
