# Changes in NEPSscaling Package Versions

For general help with the package, see:
https://www.neps-data.de/Data-Center/Overview-and-Assistance/Plausible-Values

VERSION 3.1.2

-fixed -- Issues regarding underlying package updates and changes in the SUFs

VERSION 3.1.1 

- changed -- function display_tree() changed to display_imputation_tree()

VERSION 3.0.0

- added -- NEPSshiny() function for graphical user interface
- updated -- SC2 GR w3


VERSION 2.2.0

- added -- seed for reproducibility
- added -- posterior variances of plausible values in output
- added -- standardized regression coefficients
- fixed -- NEPSscaling can now handle tibbles and haven-labelled data as bgdata
- fixed -- wrong variable name in SC3 MA w9
- fixed -- longitudinal subsample for linking now calculated from SUF wles
- fixed -- factors with more than 2 levels in bgdata are no longer treated as 
           numeric

VERSION 2.1.0

- added Starting Cohort 2: Vocabulary, Math (w9), Science (w9)
- added Starting Cohort 1: Science (w8)
- changed multiple imputation CART algorithm back to custom implementation in 
  order to provide further information about imputation process
- the output now includes the tree structures for the imputations and the 
  variable importance
- added function display_tree() to display the CART tree structure as a 
  ggplot2 graph
- added function get_imputation_names() to retrieve the names of the 
  imputations returned with the plausible values
- fixed typo in item_labels for SC3 math so that it can be used with later SUF 
  versions (SC3 version 10 and above; thanks to Felix Bittmann)


VERSION 2.0.0

- added Starting Cohort 1: Math (w5, w7), Science (w6), Cogn. Devel. (w1)
- added Starting Cohort 2: Grammar (w1, w3), Math (w2, w3, w4, w6), Science 
                           (w1, w3, w5), Native Russian/Turkish (w4),
                           Reading (w6, w9), ICT (w5), Orthography (w6)
- added Starting Cohort 3: Reading (w1, w3, w6, w9), Math (w1, w3, w5, w9),
                           ICT (w2, w5, w9), Science (w2, w5, w9),
                           Orthography (w1, w3, w5), Native Russian/Turkish 
                           (w3, w6), English (w7, w9), Scientific Thinking 
                           (w9), Listening (w6)
- added Starting Cohort 4: Reading (w2, w7, w10), Math (w1, w7, w10), ICT 
                           (w1, w7), Science (w1, w5), Native Russian/Turkish
                           (w2), English (w3, w7), Scientific Thinking (w7)
- added functions summary.pv_obj() and print.pv_obj()
- changed multiple imputation CART algorithm from custom implementation to R 
  package "mice"'s implementation, renaming the control fields for the 
  algorithm accordingly
- included argument "adjust_school_context" in plausible_values() to consider 
  higher-level variance for student cohorts: mean WLE values per school are 
  added to the background data
- linking re-scaling now bases on wle/eap/pv means instead of on eap means only
- the output now includes the means of wles, eaps and pvs as well as the number 
  of test takers included (per wave), the computation times and logged events 
  of the mice algorithm