Error in UseMethod("select_")

Question

Long time lurker, first time poster.

I'm in an introductory R course and I'm trying to create histograms and summaries for the age of diagnosis with diabetes "diabage2" and their insulin use "insulin" (Yes/No/NA). The dataset is brfss2013.

My first attempt was brfss2013 %>% group_by(insulin = "Yes") %>% summarise(MEAN = mean(brfss2013$diabage2, na.rm = TRUE), n = n())

  insulin     MEAN      n
    <chr>    <dbl>  <int>
1     Yes 51.48694 491775

Which looks fine, except I know that MEAN and n are reported for the sample mean and n, not the selected part of the sample (I've had this problem in another part of my project - not sure why it's not working. I can verify that the answer is incorrect.)

When I tried to use subset() and select for only data that met my conditions so I could easily summarise it and make histograms (i.e. one group of data where insulin = yes and one for insulin = no)

wInsulin <- subset(brfss2013, insulin = "Yes", select = c(diabage2))
woInsulin <- subset(brfss2013, insulin = "No", select = c(diabage2))

These looked the same, even though they shouldn't contain any of the same observations since they're mutually exclusive.

When I tried to use select() to trim down the set I'm using from 330 variables to three, I encountered another problem:

InsulinData <- select(brfss2013$insulin, brfss2013$diabage, brfss2013$sex, brfss2013$X_state)

gave me the error

Error in UseMethod("select_") : 
  no applicable method for 'select_' applied to an object of class "factor"

Which I have no idea what to make of.

I feel like I'm missing something very fundamental, but my lack of experience means that I don't have the foundations to understand a lot of solutions to other people's problems and the course thus far has covered more statistical theory than the actual details of dealing with R. I would really appreciate any guidance I could get.

brfss2013 doesn't come with the dplyr package, so you might want to make this more explicitly reproducible. Usually, one would write the code that grabs or loads the dataset, like library(whatever). — Frank
I'm seeing a few mistakes here, like insulin = "Yes" instead of ==; referring to cols with $ inside a dplyr chain; using subset when the dplyr idiom is filter (or some similar word). I'd recommend just browsing the R intro doc from the top and going some through examples there: cran.r-project.org/doc/manuals/r-release/R-intro.html — Frank

Brandon Bertelsen Brandon Bertelsen · Accepted Answer · 2016-08-05T19:20:18

You almost had this:

InsulinData <- select(brfss2013$insulin, 
                      brfss2013$diabage, 
                      brfss2013$sex, 
                      brfss2013$X_state)

Should be:

InsulinData <- select(brfss2013, insulin, diabage, sex, X_state)

With dplyr you only need to specify the data.frame once. select thought you were trying to select columns from the variable brfss2013$insulin, which you can't.

Also, your first set of intstructions are a bit confusing:

group_by(insulin = "yes")

You group_by(insulin) and you filter rows by filter(insulin == "yes")

Probably want something more like:

brfss2013 %>% 
  group_by(insulin) %>% 
    summarise(MEAN = mean(diabage2, na.rm = TRUE), n = n())

Error in UseMethod("select_")

1 Answers