Categorical Crime Data

SUMMARY:

The crime data frame has 9 rows and 3 columns. It contains crime status information for households in a national crime survey. crimeImpExample is an miList object, containing 10 sets of multiple imputations for the missing values in crime.

DATA DESCRIPTION:

crime contains the following columns:

ARGUMENTS:

Visit.1
the crime status of the household at the first visit of the study. "Crime-free" indicates that the household has been crime free in the previous six months. "Victim" indicates that a member of the household has been the victim of a crime in the previous six months. Missing values indicate non-response.
Visit.2
the crime status of the household at the second visit of the study. "Crime-free" indicates that the household has been crime free in the previous six months. "Victim" indicates that a member of the household has been the victim of a crime in the previous six months. Missing values indicate non-response.
count
the frequency of households with the indicated response pattern.

crimeImExample is an miList object containing 10 sets of multiple imputations; each set is a data frame containing the same variables as crime.

REFERENCES:

Schafer, Joseph L. (1997), Analysis of Incomplete Multivariate Data , Chapman and Hall, New York.

Schimert, J., Schafer, J. L., Clarkson, D. B., Fraley, C., Hesterberg, T., (2000) Analyzing Data with Missing Values in S-PLUS , Insightful Corporation, Seattle (This manual is available on-line as file Missing.pdf), Chapter 10.

SEE ALSO:

, , .

EXAMPLES:

# EM estimate 
mdLoglin(crime, frequency = count, na.proc = "em") 
#Pre-process data in order to save computation below 
crime.s <- preLoglin(crime, margins = count ~ Visit.1 : Visit.2) 
#  Fit saturated model under a Jeffreys prior using EM to get 
# starting values for DA 
crime.EM <- mdLoglin(crime.s, margins = ~ Visit.1 : Visit.2,  
                     na.proc = "em", prior = 0.5, 
                     control =  list(trace = F)) 
#start 10 independent chains from the MLE 
start.crime <- crime.EM$paramIter[rep(2, 10), ] 
crimeImpExample <- impLoglin(crime.s, prior = 0.5,  
                         start = start.crime, 
                         control = list(niter = 100, seed = 699))