Hello,
I'm new to R and work with psychometrics (Item response theory). I was using
a package called "cirt" to model a data that contains only 0s, 1s and
9s. The
package is basically used for Monte Carlo Markov Chain estimation of
parameters and I specified 5000 iterations. The model seems to run fine but
when I try to recover odds ratio statistics using the function
"irtfit" (part
of the cirt package), I keep getting the following error. I tried to already
look on R mailing list serves to understand what this could possibly mean. My
only understanding so far is that there is some entry in my data set that is
not a 0, 1 or a 9..or that there's some spacing issue? I did check my data
thoroughly though and there seems to be no issue.
Odds Ratio Statistic
item Error in rep("*", wholeWidths[i]) : invalid
'times'
argument
Can some one please advise? Any ideas would help. Thank you so much!
Rohini Sen
[[alternative HTML version deleted]]
For the list to have any hope of helping, we need at the very least a small reproducible example, one that comes with data and gives the same error. We don't even know what commands you issued. Sarah On Tue, Jul 26, 2011 at 8:48 AM, Rohini Sen <RSen at nbme.org> wrote:> Hello, > > > > I'm new to R and work with psychometrics (Item response theory). I was using > a package called "cirt" to model a data that contains only 0s, 1s and 9s. The > package is basically used for Monte Carlo Markov Chain estimation of > parameters and I specified 5000 iterations. The model seems to run fine but > when I try to recover odds ratio statistics using the function "irtfit" (part > of the cirt package), I keep getting the following error. I tried to already > look on R mailing list serves to understand what this could possibly mean. My > only understanding so far is that there is some entry in my data set that is > not a 0, 1 or a 9..or that there's some spacing issue? I did check my data > thoroughly though and there seems to be no issue. > > > > Odds Ratio Statistic > ? ? ? ? ? ? ? ? item ? Error in rep("*", wholeWidths[i]) : invalid 'times' > argument > > > > Can some one please advise? Any ideas would help. Thank you so much! > > > > Rohini Sen > >-- Sarah Goslee http://www.functionaldiversity.org
On 26/07/2011 8:48 AM, Rohini Sen wrote:> Hello, > > > > I'm new to R and work with psychometrics (Item response theory). I was using > a package called "cirt" to model a data that contains only 0s, 1s and 9s. The > package is basically used for Monte Carlo Markov Chain estimation of > parameters and I specified 5000 iterations. The model seems to run fine but > when I try to recover odds ratio statistics using the function "irtfit" (part > of the cirt package), I keep getting the following error. I tried to already > look on R mailing list serves to understand what this could possibly mean. My > only understanding so far is that there is some entry in my data set that is > not a 0, 1 or a 9..or that there's some spacing issue? I did check my data > thoroughly though and there seems to be no issue. > > > > Odds Ratio Statistic > item Error in rep("*", wholeWidths[i]) : invalid 'times' > argument > > > > Can some one please advise? Any ideas would help. Thank you so much!This is likely something only the author or maintainer of the cirt package can help you with. You can find their names in the information printed by library(help=cirt) If you need to debug it yourself, use the traceback() function to find where the error occurred, then get the source to the package from CRAN, and get to work. Debugging existing code is one way to learn a language, but it's not for everyone. Duncan Murdoch