nicegee: Lisa's GEE Regression Table Reporting Function
In lisaerein/nicereg: Suite of functions to format regression results

Description Usage Arguments

View source: R/nicegee.R

This function creates a nice looking table of regression results for a geeglm model. Input either a geeglm object or dataframe, outcome variable, vector of covariates, id/cluster variable, correlation structure, model family, and type of analysis (univariate or multiple regression). The function returns a dataframe of formatted regression results and prints the results table using kable or htmlTable.

nicegee(fit = NA, df = NA, family = NA, covs = NA, out = NA,
  id = NA, corstr = "exch", regtype = "multi", exp = NA,
  estname = NA, intercept = FALSE, refcat = FALSE, labels = NA,
  overallp = FALSE, est.dec = 2, ci.dec = 2, pval.dec = 3,
  color = "#EEEEEE", kable = TRUE, htmlTable = FALSE, title = "",
  footer = TRUE)

`fit`	geeglm model object [fit or family/covs/out/id/corstr are REQUIRED].
`df`	Dataframe [fit or family/covs/out/id/corstr REQUIRED].
`family`	Character. Model family name in quotes ("guassian", "binomial", "poisson") [fit or family/covs/out/id/corstr are REQUIRED].
`covs`	Character. Vector of covariates to include in model [fit or family/covs/out/id/corstr are REQUIRED].
`out`	Character. Outcome for regression model [fit or family/covs/out/id/corstr are REQUIRED].
`id`	Character. Name of ID/cluster variable [fit or family/covs/out/id/corstr are REQUIRED].
`corstr`	Character. Name of correlation structure. Default is "exch".
`regtype`	Logical. Should the covariates be run separately ("uni") or together in a multiple regression model ("multi") [fit or family/covs/out/id/corstr are REQUIRED].
`exp`	Logical. Option to exponentiate coefficients and CI's. Default is NA (estimates exponentiated for binomial and poisson models by default).
`estname`	Character. Label for regression estimate. Default is NA (program will choose a reasonable name depending on model type).
`intercept`	Logical. If TRUE the intercept will be included in the table (muliple regression only). Default is FALSE.
`refcat`	Logical. If TRUE the table will create a separate line for the reference category. Default is FALSE.
`labels`	Character vector. Covariate labels in same order as covs. Default is NA (variable names are used).
`overallp`	Logical. If TRUE, a likelihood ratio test pvalue will be calculated for each variable (via anova, Chisq test). Default is FALSE.
`est.dec`	Numeric. Number of decimal places for estimates. Default is 2.
`ci.dec`	Numeric. Number of decimal places for 95 percent confidence interval. Default is 2.
`pval.dec`	Numeric. Number of decimal places for pvalues. Must be an integer from 1 to 4. Default is 3.
`color`	Character. Color to use for htmlTable striping. Default is "#EEEEEE" (light grey). Use "white" for no striping.
`kable`	Logical. If TRUE, table will be formatted using the kable packge. Default is TRUE.
`htmlTable`	Logical. If TRUE, the table will be printed using htmlTable. Default is FALSE.
`title`	Character. Optional title above table. Default is "".
`footer`	Logical. If TRUE, table will include a footnote with model details, nobs, R2. Default is TRUE.