R to create a new variable, which is the mean of other several variables

Question

I am a beginner of R, recently I met some troubles in creating a new variable with mutate() function. The data frame "evs" has six variables:v1,v2,v3,v4,v5,v6. I hope to calculate their average and assign the average to a new variable "intp.trust" ,so I use the following commands:

evs <- mutate(evs,intp.trust= mean("v1","v2","v3","v4","v5","v6"))

But the new varriable "intp.trust" turns to have only NA. Is there anything wrong with my command? I hope someone can help me with it.

Use `evs %>% mutate(intp.trust = rowMeans(across(v1:v6), na.rm = TRUE)))` — akrun, Oct 15 '21 at 23:23

score 1 · Answer 1 · answered Oct 15 '21 at 23:31

1

Use rowMeans as it is vectorized

library(dplyr)
evs <- evs %>% 
   mutate(intp.trust = rowMeans(across(v1:v6), na.rm = TRUE))

answered Oct 15 '21 at 23:31

akrun

874,273
37
540
662

1

Hi thank you, I'll try it! – Xingchen LIU Oct 15 '21 at 23:35

score 0 · Answer 2 · answered Oct 15 '21 at 23:29

Edit: Corrected, I missed out the required c

I believe mean doesn't work over multiple vectors, so you need to tell dplyr to do the operation row-by-row. You can do this by using the "rowwise" function, i.e.:

evs <- evs %>% rowwise() %>%
  mutate(intp.trust = mean(c(v1, v2, v3, v4, v5, v6)))

cgvoller · Accepted Answer · 2021-10-15T23:41:23.410

0

In dplyr

evs %>% rowwise() %>% mutate(intp.trust = mean(v1:v4))

Another solution:

evs$intp.trust <- apply(evs[1:6], 1, mean, na.rm = TRUE)

Output:

      v1       v2       v3       v4 intp.trust
1 3.386651 7.651824 3.537067 7.490539   5.516520
2 9.749669 6.686065 9.328821 1.676758   6.860328
3 1.231070 4.339285 9.274955 1.046044   3.972839

edited Oct 15 '21 at 23:41

answered Oct 15 '21 at 23:30

cgvoller

873
1
14

R to create a new variable, which is the mean of other several variables

3 Answers3