Dear Researchers, I am new to R programming, and I need some help. Let's say I want to calculate the mean duration of the vowel "i" only for all speakers and all settings (region: rural & urban; control: yes & no) what command should I use?
What if I want to calculate the mean if "i" for all participants in (region: rural) only or (control: no) for example?
My understanding from your first question is you want to calculate mean for "i" including data from all speakers, region and control.
install.package(dplyr) # install package
library(dplyr) #attach package for use
your_data%>% #your dataframe
subset(vowel=="i")%>% #Use subset to obtain only data for "i"
summarise(mean=mean(vowel), SD=sd(vowel)) # calculate mean and sd
Multiple apporaches to your second question:
You can use the same approach above adding more conditions in subset()
your_data%>% #your dataframe
subset(vowel=="i" & region =="rural" & control == "no")%>%
summarise(mean=mean(vowel), SD=sd(vowel))
Or calculate means for all the groups you want using group_by() and obtain the values as when needed.
your_data%>% #your dataframe
group_by(region, control, vowel)%>% #group data by regionm control and vowels
summarise(mean=mean(vowel), SD=sd(vowel))
It's much simpler than Yin Ying Chin suggests, no need to use any add-on package:
I suggest first to subset only the rows containg data for the vovel "i":
df
Dear @Dan Li,
You may also try following codes
library (dplyr)
dplyr::filter (data=your.data, your.data$vowel =="i") %>%
group_by (region, control) %>%
summarise (mean_duration =mean(duration))
1) It will be worth your while to learn a little of the R language. Jochen Wilhelm has given you some suggests using the tapply() function in the base R package. Once you get familiar with some common R functions, you will find that manipulating data is easier than e.g. using a spreadsheet. I use R all the time now, even in my personal life, for little tasks that others might use a calculator for, or in cases where I might need to repeat calculations in the future.
If I were approaching this problem, I would probably use the brackets for subsetting:
mean(Data$duration[Data$vowel=="i"]
2) There are functions in add-on packages that make some things easier. There are a few caveats with using add on packages *.
In the case of this problem, I might use:
library(FSA)
Summarize(duration ~ vowel, data=Data)
You can easily drill down more:
mean(Data$duration[Data$vowel=="i" & Data$region=="rural"]
library(FSA)
Summarize(duration ~ vowel + region, data=Data)
With the caveat that I wrote it, I have other options for summarizing data here:
https://rcompanion.org/handbook/C_02.html
# # #
* Some add-on R packages are well-established. a) But others may be less so. And it's important to remember that anyone can write and publish an R package. It's a good idea to double check the results for a function from a package that may not be well-established. b) Also, authors may change how a function works or may deprecate a function. This can cause challenges when you rely on a specific function in a package, and then that function changes or is deprecated, especially if you have older code snippets saved. c) On that note, even the base/primary R packages may change occasionally. Recently, R essentially changed the default behavior of read.table() from stringsAsFactors=TRUE to stringsAsFactors=FALSE. This caused a headache for me, necessitating changing code I had saved or posted.
Using the tidyverse, and assuming your data is stored in "df" and you are tying to calculate the mean for each speaker, by region and control, you could use something like:
df %>%
filter(vowel == "i") %>%
group_by(speaker, region, control) %>%
summarise(mean_duration = mean(duration))
For your second case, change slightly the code above:
df %>%
filter(vowel == "i", region == "rural", control == "no") %>%
group_by(speaker) %>%
summarise(mean_duration = mean(duration))
Agree to the crowd. There are many packages available to suit your preferences (i.e., avoid creating new objects in the R environment/keeping code lines short) and which one you find easier to make sense of.
For reasons Sal Mangiafico mentioned, good to know of the alternatives available and learning basic R syntax would be of huge help.
Free R resource broken into 15-20 minute learning blocks.
https://www.tutorialspoint.com/r/r_mean_median_mode.htm
All the best.
- Badges
- Science method