ordinalRR (version 1.1)

preprocess: Format an ordinal R&R data frame into object required by function ordinalRR.

Description

This function accepts a concise data frame containing the ordinal responses from an R&R study and expands it into the necessary multinomial counts, so it can be an input to function ordinalRR() for Bayesian data analysis.

Usage

preprocess(x, J=3, K=2, H=4)

Arguments

x

‘data.frame’ containing J*K columns and entries from the H-point ordinal scale 1:H. The required format is a row for each part and blocks of K adjacent columns for the repetitions of each of J raters, e.g., rater 1's columns are the first K and rater J's columns are the last K.

J

‘positive integer’ defining the number of raters.

K

‘positive integer’ defining the number of repetitions per rater.

H

‘positive integer’ defining the H-point ordinal scale on {1,...,H}.

Value

I

‘positive integer’ number of parts.

J

‘positive integer’ number of raters.

K

‘positive integer’ number of repetitions per rater.

H

‘positive integer’ length of the H-point ordinal scale on {1,...,H}.

x

‘data.frame’ containing J*K columns and entries from the H-point ordinal scale 1:H. Each part is a row, and there are blocks of K adjacent columns for the repetitions of each of J raters, e.g., rater 1's columns are the first K and rater J's columns are the last K.

R

‘array’ x is expanded into a 3-dimensional array (i.e., part 1:I, operator 1:J, ordinal value 1:H) with multinomial counts.

preprocess

‘Boolean’ will be TRUE if the data are ready for input into function ordinaRR() for Bayesian analysis with JAGS.

See Also

followup

Examples

# NOT RUN {
data(followup)
followup
preprocess(followup) 
# }