TukeyHSD() function in R-CodePudding

I am trying to run TukeyHSD for a data set (40 observations). The outcome is 'mbvs' and the two factors are sex (0 or 1) and treatment (0 or 1) ?

I am trying to find out what is the pairwise difference in the mean outcome between treatment and control groups in men (sex=0)?

Could someone help me with the code for this ? Thank you.

CodePudding user response：

Look, I do not whether this answer can help you, but I've tried to semplify the dataset you provided with a smallest one, just to make you understand the logic presiding over it.

library(multcomp)
library(dplyr) 

mbvs <- c(6.4, 6.2, 6.9, 6.9, 5.4, 8.4)
sex <- c(0,0,0,0,0,1)
tmt <- c(0,1,0,0,0,1) 


   
#let's make this as your dataset 

df1 <- data.frame(mbvs, sex, tmt, stringsAsFactors = FALSE)

postHocs = df1 %>% 
  mutate(sex = factor(sex, levels = c('0', '1')), 
         tmt = factor(tmt, levels = c('0', '1'))) %>% 
  filter(sex == '0') %>%
  aov(mbvs ~ tmt, data = .) %>% 
  glht(., linfct = mcp(tmt = 'Tukey'))

The final result I've obtained is

     General Linear Hypotheses

Multiple Comparisons of Means: Tukey Contrasts


Linear Hypotheses:
           Estimate
1 - 0 == 0     -0.2

Alternatively, with the TukeyHSD function you can run this code

postHocs = df1 %>% 
  mutate(sex = factor(sex, levels = c('0', '1')), 
         tmt = factor(tmt, levels = c('0', '1'))) %>% 
  filter(sex == '0') %>%
  aov(mbvs ~ tmt, data = .) %>% 
  TukeyHSD(., conf.level=.95)

  Tukey multiple comparisons of means
    95% family-wise confidence level
Fit: aov(formula = mbvs ~ tmt, data = .)

$tmt
    diff       lwr      upr     p adj
1-0 -0.2 -2.715945 2.315945 0.8166266