In meghapsimatrix/MeghaMCSE: Performance criteria and monte carlo simulation error

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

MeghaMCSE

The goal of MeghaMCSE is to caclulate performance criteria measures such as bias, relative bias, mean squared error, root mean squared error etc. for results from simulation studies. In addition to calculating the performance measures, the package also calculates associated Monte Carlo Standard Errors.

Installation

You can install the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("meghapsimatrix/MeghaMCSE")

Example

This is a basic example which shows you how to solve a common problem:

library(MeghaMCSE)

library(tidyverse)
library(broom)
library(kableExtra)

set.seed(20191228)

# function to create normally distributed data for each group to run t test

generate_dat <- function(n = 50, effect_x){

  dat <- tibble(group_1 = rnorm(n, 0, 1), 
                group_2 = rnorm(n, effect_x, 1))

  return(dat)

}


# function to calculate t-test, pulls out estimate of the mean difference, p val and ci 

estimate_t <- function(sim_dat){

  res <- tidy(t.test(sim_dat$group_2, sim_dat$group_1)) %>%
    select(estimate, p_val = p.value, ci_low = conf.low, ci_high = conf.high)

  return(res)

}


# generating 1000 iterations

results <- 
  rerun(1000, {
    dat <- generate_dat(effect_x = .5)
    estimate_t(dat)
  }) %>%
  bind_rows()


# running calc_mcse

calc_mcse(estimates = results$estimate, true_param = .5, K = nrow(results), perfm_criteria = c("bias", "variance", "mse", "rmse", "relative bias", "relative mse"))


calc_mcse(estimates = results$p_val, true_param = .5, alpha = .05, K = nrow(results), lower_bound = results$ci_low, upper_bound = results$ci_high, perfm_criteria = c("rejection rate", "coverage", "width"))