tidycomm 0.4.1

tidycomm 0.3.0

New features

• Converted tibble returns into tdcmm/tibble return objects (they behave just like tibbles but are in essence our own objects now)
• Added partial correlation in correlate(..., partial = z_var)
• Added correlation with a focus variable correlate(..., with = focus_var)
• Added linear regression regress()
• Added one-sample t-test t_test(..., mu = ...)
• Added reverse_scale(), minmax_scale(), z_scale(), center_scale(), setna_scale(), recode_cat_scale(), recode_scale(), and dummify_cale() to shift and modify continuous and categorical scales
• Added tab_percentiles()
• Added visualize() to visualize almost everything
• Added snscomments and incvlcomments as additional data sets

Minor changes

• Changed unianova() and t_test() to build on leveneTest()
• Allowed test_icr() to work with grouped data
• Converted all code examples in documentation to be built on tidycomm-provided data sets
• Added omega_squared, Levene_p, and var_equal columns to default return from unianova()
• Added d, se, t, and df
• Removed null.value from list of return values in unianova() post-hoc test
• Renamed unianova() return column names to Variable (previously: Var), Group_Var (prev. term), Delta_M (prev. estimate), p (prev. adj.p.value), conf.lower and conf.upper (prev. conf.low and conf.high)

Bugfixes

• Unified output to coherent number of after-comma digits

tidycomm 0.2.0

New features

• Added Fretwurst’s Lotus and S-Lotus intercoder reliability coeffecients to test_icr() function
• describe_cat() function added to describe categorical variables

Minor changes

• More descriptive error messages for common errors
• describe() now also reports 95% confidence intervals
• describe() now reports valid N instead of full N
• test_icr() now works with tidyselect selection

Bugfixes

• Empty groups are dropped if describing with more than one grouping variable
• Krippendorff’s Alpha returns 1 if variable has only one category
• unianova() now works with variable names containing whitespace
• Groups are dropped for test_icr() to avoid computational issues