litnum1215: Plausible values imputation using the PIAAC 2012 and PIAAC-L...
In jcgaasch/PVPIAACL: Plausible Values estimation with the PIAAC-L data

Description Usage Arguments Value References

This function estimates a four-dimensional latent regression item response model (first dimension: PIAAC 2012 literacy, second dimension: PIAAC 2012 numeracy, third dimension: PIAAC-L 2015 literacy, fourth dimension: PIAAC-L 2015 numeracy) for binary item response data considering partially missing covariate data. For more detailed information on the statistical model and the estimation algorithm, see the PIAAC-L technical report on scaling (Carstensen, Gaasch & Rothaug, 2017).

1	litnum1215(path, X = NULL, nopvs = 10, itermcmc = 22000, burnin = 2000)

`path`	full path of the folder containing the data files ZA5845 and ZA5989_Persons_15 in Stata format.
`X`	data frame containing the sequential ID (named `seqid`) and background variable from the PIAAC and PIAAC-L Scientific Use Files. They can be numeric or factor variables and contain missing values coded as `NA`. With `X` set to `NULL` (default) an empty population model is estimated.
`nopvs`	number of plausible values to draw for each respondent.
`itermcmc`	number of MCMC iterations.
`burnin`	number of burnin iterations.

list with nopvs elements, each containing a data frame of the sequential ID, plausible values for each dimension and, if specified, imputed versions of X. Resulting plausible values are transformed onto the PIAAC 2012 scale (weighted means and standard deviations based on the SUF). Additionally each list element is saved as a Stata file in the folder specified by path.

Carstensen, C. H., Gaasch, J.-C., & Rothaug, E. (2017). Scaling PIAAC-L cognitive data: technical report. Manuscript in preparation.

jcgaasch/PVPIAACL documentation built on May 12, 2019, 8:43 a.m.