Changelog • regdatasets

regdatasets 0.2.0

naep dataset removed: The NAEP 2005 Grade 8 Mathematics dataset (17,606 x 301) has been removed because respondent-level NAEP microdata with weights are restricted-use under NCES data licensing policy. The package now contains 24 datasets. For NAEP analyses, use the EdSurvey R package to access NAEP data directly from NCES.
berkeley$departme renamed to department: The truncated Stata 8-character variable name has been corrected. Code using berkeley$departme will need to be updated.
gcse reduced from 9 to 6 columns: Removed u0, u1, and filter__ columns, which contained Stata multilevel model artifacts (placeholder values of -1e30). Remaining columns: school, student, gcse, lrt, gender, pred.

Sentinel missing values recoded to NA in disc (variables bys34a, bys34b: codes 8 and 98) and nels_data (82 sentinel values across multiple variables: codes 99.98, 99.99, 99.998, 998, 100).
titanic$embarked: Two blank strings converted to NA.
disc$race and disc2$race: Converted from integer codes (1-4) to unordered factor with levels: Asian, Hispanic, Black, White.

list_datasets() function: Returns a tibble summarizing all 24 datasets with name, dimensions, chapter mapping, and description.
LICENSE.note: New file documenting data source provenance and redistribution basis for each dataset.
inst/CITATION: Added so citation("regdatasets") returns a proper bibliographic entry.

Ethical context sections added to penalty (racial disparities in capital sentencing, Radelet 1981, McCleskey v. Kemp), disc and disc2 (school-to-prison pipeline, race in school discipline), and titanic (note of respect for individuals represented).
Cross-references (@seealso) added linking related dataset pairs: berk_sub/berkeley, disc/disc2, hsb_sub/hsbs1, plus womenlf/carData::Womenlf overlap disclosure and berkeley/datasets::UCBAdmissions comparison.
gcse source citation completed: Goldstein, H., Rasbash, J., Yang, M., et al. (1993). Oxford Review of Education, 19(4), 425-433.
Updated disc/disc2 documentation to reflect sentinel value cleaning and race factor conversion.

Added tibble to Imports in DESCRIPTION to ensure correct deserialization of tibble objects.
Expanded test suite: From 73 lines / 4 tests to comprehensive coverage of all 24 datasets (tibble class, dimensions, column names, factor levels, sentinel value absence, binary variable validation, Stata artifact absence).
Added NEWS.md (this file).