R, Simulation, p-value, histogram - r

Question

library(tidyverse)
library(broom)
library(dplyr)
# create a tibble with an id column for each simulation and x wrapped in list()
sim <- tibble(id = 1:1000,
x = list(rbinom(1000,1,0.5))) %>%
# to generate z, pr, y, k use map and map2 from the purrr package to loop over the list column x
# `~ ... ` is similar to `function(.x) {...}`
# `.x` represents the variable you are using map on
mutate(z = map(x, ~ log(1.3) * .x),
pr = map(z, ~ 1 / (1 + exp(-.x))),
y = map(pr, ~ rbinom(1000, 1, .x)),
k = map2(x, y, ~ glm(.y ~ .x, family="binomial")),
# use broom::tidy to get the model summary in form of a tibble
sum = map(k, broom::tidy)) %>%
# select id and sum and unnest the tibbles
select(id, sum) %>%
unnest(cols = c(sum)) %>%
# drop the intercepts and every .x with a p < 0.05
filter(term !="(Intercept)",
p.value < 0.05)
sim
j=exp(sim %>% select("estimate"))
OR=as.numeric(unlist(j))
mean(OR)
hist(OR,main=NULL,freq=T,breaks=10)
abline(v=mean(OR),lwd=4,col=1)
The question here: now I extract all the value which p<0.05, now I using the code "hist(OR,main=NULL,freq=T,breaks=10)" to make a histogram for the odds ratio. The new thing I want to do is make another histogram(like without any condition for p-value) overlapping the original one, then I could compare the histogram with the different p-value in one plot, which code can work with that?

Answer 1

This solution repeats the question's code but
stop the pipe right after unnest(cols = c(sum));
create a simOR like you have continued the pipe and a simAll but this time not filtering the p-values.
First the question's code. Note that if package tidyverse is loaded there is no need to load package dplyr.
I also set the RNG seed to make the results reproducible.
library(tidyverse)
library(broom)
# create a tibble with an id column for each simulation and x wrapped in list()
set.seed(2020)
sim <- tibble(id = 1:1000,
x = list(rbinom(1000,1,0.5))) %>%
# to generate z, pr, y, k use map and map2 from the purrr package to loop over the list column x
# `~ ... ` is similar to `function(.x) {...}`
# `.x` represents the variable you are using map on
mutate(z = map(x, ~ log(1.3) * .x),
pr = map(z, ~ 1 / (1 + exp(-.x))),
y = map(pr, ~ rbinom(1000, 1, .x)),
k = map2(x, y, ~ glm(.y ~ .x, family="binomial")),
# use broom::tidy to get the model summary in form of a tibble
sum = map(k, broom::tidy)) %>%
# select id and sum and unnest the tibbles
select(id, sum) %>%
unnest(cols = c(sum))
Now create the two data sets to be plotted.
simOR <- sim %>%
# drop the intercepts and every .x with a p < 0.05
filter(term !="(Intercept)", p.value < 0.05)
j <- exp(simOR %>% select("estimate"))
OR <- as.numeric(unlist(j))
mean(OR)
And the data set with all the rows, dropping only the intercepts.
simAll <- sim %>%
filter(term !="(Intercept)")
j <- exp(simAll %>% select("estimate"))
All <- as.numeric(unlist(j))
mean(All)
Now plot the histograms (not overlapped).
op <- par(mfrow = c(2, 1))
hist(OR, main = NULL, freq = TRUE, breaks = 10)
abline(v = mean(OR), lwd = 4, col = 1)
hist(All, main = NULL, freq = TRUE, breaks = 10)
abline(v = mean(All), lwd = 4, col = 1)
par(op)