- df, benchmark
A data frame containing the set of respondents or
benchmark set of respondents to compare, or a character string containing the
name of the set of respondents or benchmark set of respondents. All independent
and dependent variables must be inside both data frames.
- independent
A list of strings containing the independent variables (x)
for comparison. Every independent variable will be used in every model to
estimate the dependent variable (y). When a formula_list is provided,
independent will be ignored.
- dependent
A list of strings containing the dependent variables (y) for
comparison. One model will be computed for every dependent variable (y)
provided. When a formula_list is provided, dependent will be
ignored.
- formula_list
A list of formulas to use in the regression models. If
given, dependent and independent parameters will be ignored.
- family
A family input, that can be given to glm or
svyglm. Additionally, if "ols" is given,
gaussian(link = "identity"), and if "logit" is given,
binomial(link = "logit") is used.
- rm_na
A character to determine how to handle missing values. For this two
options are supported. If rm_na = "pairwise" NAs will be removed
separately for every model. Only cases containing NA on one of the variables used
in the respective model will be removed (all independent variables but only
the respective dependent variable). If rm_na = "listwise" all cases containing NA
on one of the dependent or independent variables are removed.
- out_output_list
A logical value. If out_output_list = TRUE, a
list will be returned, containing the separate interaction models calculated
with the glm function or svyglm
in case of weighting, as well as a summary object for every model.
Standard errors and p-values of these models are always calculated without
robustness methods.
- out_df
If TRUE, the used data frames will also be part of the output list.
- out_models
If True, GLM model objects will be part of the returned object.
- print_p
If TRUE, in addition to the difference in Average Discrete Change (ADC),
p-values will be printed.
- print_se
If TRUE, additionally standard errors will be printed.
- weight, weight_bench
A character vector containing the name of the weight
variable in the respective data frame. If provided the data frame will be weighted
using the svydesign function. Also id must be provided.
- id, id_bench
A character vector containing the name of the id variable in the respectiv
data frame. Only needed for weighting.
- strata, strata_bench
A character vector containing the name of the strata variable
in the respective data frame. It is used in the svydesign
function for weighting.
- nest, nest_bench
A logical vector that is used in the svydesign
function for the respective data frame.
- robust_se
A logical value If TRUE instead of normal standard errors,
heteroscedasticity-consistent standard errors will be used in the analysis to
calculate them the vcovHC and coeftest
packages are used.
- p_adjust
A logical input or character string indicating an adjustment
method usable in the method parameter of p.adjust.
If set to TRUE the Bonferroni adjusted p-values are used in inference.
- names_df_benchmark
A vector containing first the name of df and benchmark.
- silence_summary
A logical value, to indicate if the
printed summary should not be printed instead.
- nboots
A numeric value indicating the number of bootstrap replications.
If nboots = 0 no bootstrapping will be performed. Else nboots must
be >2. Note, that bootstrapping can be very computationaly heavy and can
therefore take a while.
- boot_all
If TURE, both, dfs and benchmarks will be bootstrapped. Otherwise
the benchmark estimate is assumed to be constant.
- parallel
If TRUE, all detected cores will be used in bootstrapping.
- adjustment_vars
Variables used to adjust the survey when using raking
or post-stratification.
- raking_targets
A List of raking targets that can be given to the rake
function of rake, to rake the df.
- post_targets
A List of post_stratification targets that can be given to the rake
function of postStratify, to post_stratificatify the df.
- percentile_ci
If TURE, cofidence intervals will be calculated using the percentile method.
If False, they will be calculated using the normal method.