Movatterモバイル変換

resquin 0.1.1

Changed the way the mahalanobis distance is calculcated inresp_distributions(), if missing values are allowed. Nowwithin respondent mean imputation is used. Before missing values wereturned to a value of 0, which is wrong in almost any case and would haveskewed the results of respondents with missing values. Mean imputationis not an ideal solution, but it allows the observed data to influencethe value of the mahalanobis distance value under missing data. Becausewithin respondent mean imputation is not ideal, a new section in thedescription is added to call for caution when interpreting themahalanobis distance values produced byresp_distributions() ifmin_valid_responses issmaller than 1.
Polished documentation (fixing typos etc.)
Added vignette on flagging respondents withflag_resp().
Added a disclaimer for handling of missing data inresp_nondifferentiation.
Added documentation on s3 methods.
Fixed bug in s3 print function which would crash ifresp_styles() was used on even numbered responsescales.
Added tests for functions added since 0.0.2.
Now depends on R version being >= to R version 4.1.
Added example data setnep from the GESIS panel.

Addedresp_patterns() andresp_nondifferentiation() as a new function.
Addedid column to all outputs to make it easier toidentify respondents or merge function outputs to data frames.id is eitherTrue for an integer id,False for noid column, or a vector of uniqueinteger or character values identifying each respondent.
Addedflag_resp() function to quickly create andcompare different flagging strategies based on response qualityindicators.
Added s3 types to outputs ofresp_*() andflag_resp() functions.
Added s3 print, summary and plot methods for outputs ofresp_*() functions.
Added s3 summary method forflag_resp() output.

Changes in response to CRAN team. Changed license description inDESCRIPTION and removed License.md
Removed warning on development version at the start of thereadme.md