Changes in version 2.0.0 (2025-08-26) - Fixed issue if parameter period was type haven::labelled in calc.stError(). - Refactored code to make use of new data.table variable env. Makes code more readable and stable. - calc.stError() has new parameter group.diff. If group.diff=TRUE differences between values defined in group will also be calculated. For instance library(surveysd) set.seed(1234) eusilc <- demo.eusilc(n = 4,prettyNames = TRUE) dat_boot <- draw.bootstrap(eusilc, REP = 3, hid = "hid", weights = "pWeight", strata = "region", period = "year") dat_boot_calib <- recalib(dat_boot, conP.var = "gender", conH.var = "region") # estimate weightedRatio for povertyRisk per period err.est <- calc.stError(dat_boot_calib, var = "povertyRisk", fun = weightedRatio, group = "gender", group.diff = TRUE) will produce difference in poverty rate between male and female, in addition to other estimates. - Updated vignette accordingly. - Fixed bug with parameter adjust.var and fun.adjust.var in function calc.stError(). - Changed parameter national to relative.share in calc.stError() included a deprecate-message if national = TRUE is supplied. - Fixed issue with parameter looseH in function ipf(). - Fixed bug that variables names created by internal functions where not properly cleaned up before returning data.table output. - Improved summary.ipf() and included print.summary.ipf() in order to make the output of ipf() as well as convergence issues when using ipf() more intuitive. - New parameter method in draw.bootstrap() and rescaled.bootstrap(). Can bei either "Preston" or "Rao-Wu", see ?draw.bootstrap() or the new vignette. - draw.bootstrap() can now be used to draw bootstrap replicates if bootstrap replicates where drawn for previous year with parameter already.selected. already.selected expects a list of data.tables indicating if a record was already included in a bootstrap replicates in the previous period - New function get.selection() to create input for parameter already.selected in draw.bootstrap(). - Updated and included (more) unit tests. Changes in version 1.3.2 - rescaled.bootstrap() has additional parameter period which is identical to the one in draw.bootstrap. If period is not NULL the boostraps will be drawn such that in each period and strata/cluster only $\floor{\frac{n}{2}}$ records are drawn. This produces more consisten results and should make calibration afterwards easier. - Improved numerical weighting with computeLinearG1 for use with ipf(). computeLinearG1 is now more stable when only numerical variables are used for weighting. - parameter numericalWeighting can be passed to recalib and will correctly be passed along to function ipf Changes in version 1.3 - new parameter minMaxTrim in ipf() to trim weights - fix bug for numericalWeighting reported "converged" when it did not converge - fix bug when checking cluster and strata for bootstrapping Changes in version 1.2 - Add vignette about ipf() - Resolve bug that appeared if constraints contained only one household or only one person (#17) Changes in version 1.1.1 - recalib() accepts conP and conH in the same way as ipf - recalib() has arguments epsP and epsH, to make convergence limits more transparant - fixed bug in ipf() when supplying hid="hid", see https://github.com/statistikat/surveysd/pull/20. Thanks @asiripanich Changes in version 1.1.0 (2019-07-15) - uncouple surveysd from simPop - use demo.eusilc() for all examples and unit tests in this package - remove simPop from the "Suggests" field (DESCRIPTION) - remove hardcoded variable names in default parameters - allow calls to recalib() without specifying conP.vars or conH.vars Changes in version 1.0.2 - fix bug with NSE - update documentation for calc.stError Changes in version 1.0.1 - fix bug with fpc calculation (0824834) - fix bug with dummy data (5adb6b7) - improve readability of code and resolve linters - resolve issues when datasets have columns with certain names (#7, #10) - automatize the gh-pages builds - css updates - automated linter-checking - new badges (code coverage, cran) - update setup for vignettes - fix issue when household column is not a factor (#12) Changes in version 1.0.0 (2019-01-25) - copy simPop::ipu2() and some related functions to surveysd. - add bugfixes for ported functions - simplify tests - update references for the methodology vignette - fix sampling with full population in some strata Changes in version 0.2.3 - Fix citation - Resolve issue that occured when columns had multiple classes Changes in version 0.2.2 - Update some default parameters (#4, #5) - Split README into README + Get Started - Bugfixes - Update documentation titles, function categories and vignette titles Changes in version 0.2.1 - Create vignette for error estimation - Bugfixes for plotting and ... argument in calc.stError - Documentation updates. Make example plots available in gh-pages Changes in version 0.2.0 - Automatically pass variable names for household id, weights and reference period from draw.bootstrap to downstream functions - Reorganize names of R source files Documentation updates - Use markown pre-processing for documentation - Add github pages, more badges and a logo - export draw.bootstrap and include prettyNames argument. update documenentation examples to use new variable names - Don't load laeken and data.table in examples - rewrite vignettes/TheoryWord.Rmd into a dynamic markdown file describing the methodology of the package