Influence of p-value cutoff change

pvalue.r

library(tidypvals)
library(dplyr)
library(ggplot2)
aj1 = anti_join(head2015,chavalarias2016)
aj2 = anti_join(chavalarias2016,head2015)
sj1 = semi_join(head2015,chavalarias2016)
allp = rbind(aj1,aj2,sj1)

allp = rbind(allp,brodeur2016)

allp0 = allp %>% 
        filter(!is.na(field)) %>%
        group_by(field) %>% 
        mutate(p0.05 = (pvalue<0.05)) %>%
        mutate(p0.005 = (pvalue<0.005)) %>%
        summarise(avg1 = mean(p0.05),avg2 = mean(p0.005)) %>%
        mutate(wrong = avg1-avg2)

allp0 %>%
        ggplot(aes(field, wrong)) +
        geom_col(show.legend = FALSE) +
        ylab("percentage of not significant paper if p-value cutoff is 0.005") +
        coord_flip()

allp0 %>%
        ggplot(aes(field, avg1)) +
        geom_col(show.legend = FALSE) +
        ylab("percentage of significant paper if p-value cutoff is 0.05") +
        coord_flip()

allp0 %>%
        ggplot(aes(field, avg2)) +
        geom_col(show.legend = FALSE) +
        ylab("percentage of significant paper if p-value cutoff is 0.005") +
        coord_flip()