Odd behavior from add.design.data

posts related to the RMark library, which may not be of general interest to users of 'classic' MARK

Odd behavior from add.design.data

Postby JeffHostetler » Fri Jun 03, 2011 2:23 pm

I'm getting odd results when I use add.design.data with pim.type='time'. This:
Code: Select all
GvolCJS.proc=process.data(GvolCJS,model="CJS",begin.time=1,groups=c("SEX","SITE"),time.intervals=c(3.5,3.25,3,9.75,3.5,3,3.25,3.5,4.25,2.5,4.25,2,3.25,3,1.75,1.5,1.5,1.75,1.5,1.5,1.25,2,1.25,1.5,1.25,2.75,2.25,1,2,2.25,2.25,5.25,2.5,0.75,0.75,1.25,2.25,1),nocc=39)
GvolCJS.ddl=make.design.data(GvolCJS.proc, parameters=list(Phi=list(pim.type='time'),p=list(pim.type='time')))
GvolCJS.ddl=add.design.data(GvolCJS.proc,GvolCJS.ddl,parameter="p",type="time",bins=c(1,90.75,96),name="capType",right=FALSE)
with(GvolCJS.ddl$p, table(capType, time))

produces this:
Code: Select all
         time
capType   4.5 7.75 10.75 20.5 24 27 30.25 33.75 38 40.5 44.75 46.75 50 53 54.75 56.25 57.75 59.5 61 62.5 63.75 65.75 67 68.5
  [1,90]   15   15    16   16 15 13    13    13 13   14    16    15 14 13    12    12    12   13 12   13    13    14 14   15
  (90,96]   1    1     0    0  1  3     3     3  3    2     0     1  2  3     4     4     4    3  4    3     3     2  2    1
         time
capType   69.75 72.5 74.75 75.75 77.75 80 82.25 87.5 90 90.75 91.5 92.75 95 96
  [1,90]     15   15    14    12    13 13    11   11 11     9    9    10 11 13
  (90,96]     1    1     2     4     3  3     5    5  5     7    7     6  5  3


If I don't constrain the PIM types, it works fine:
Code: Select all
           time
capType     4.5 7.75 10.75 20.5  24  27 30.25 33.75  38 40.5 44.75 46.75  50  53 54.75 56.25 57.75 59.5  61 62.5 63.75 65.75  67
  [1,90.8)   16   32    48   64  80  96   112   128 144  160   176   192 208 224   240   256   272  288 304  320   336   352 368
  [90.8,96]   0    0     0    0   0   0     0     0   0    0     0     0   0   0     0     0     0    0   0    0     0     0   0
           time
capType     68.5 69.75 72.5 74.75 75.75 77.75  80 82.25 87.5  90 90.75 91.5 92.75  95  96
  [1,90.8)   384   400  416   432   448   464 480   496  512 528     0    0     0   0   0
  [90.8,96]    0     0    0     0     0     0   0     0    0   0   544  560   576 592 608


I can also get the results I want by setting up the capType variable manually:
Code: Select all
GvolCJS.ddl$p$capType2 = 'basic'
GvolCJS.ddl$p$capType2[GvolCJS.ddl$p$time %in% c(90.75,91.5,92.75,95,96)] = 'tree'
GvolCJS.ddl$p$capType2=factor(GvolCJS.ddl$p$capType2)


Was I wrong to constrain the PIM types? I thought pim.type='time' was okay if one doesn't have any age or cohort covariates.

I'm using RMark version 2.0.0, in case that matters.

Thanks,

Jeff
JeffHostetler
 
Posts: 13
Joined: Thu Sep 18, 2008 11:31 am
Location: Smithsonian Institution

Return to RMark

Who is online

Users browsing this forum: No registered users and 0 guests

cron