Comparing means of categorical data

[SL_Mabaso]

I am trying to compare means of categorical data, that is education (integer) grouped by sex (Male and Female). So I have used the following code to view their means:

gss %>%
group_by(sex) %>%
  summarise(xbar = mean(educ))

This code provides a table:

Sex xbar
Male NA
Female NA

I dont get why this is the case since my education (educ) vector has numerical values and NA (very few).

[gueyenono]

Hi @SL_Mabaso

The result would be the same even if there was a single NA value in each category. If you want to ignore those values, then you need to add the na.rm = TRUEargument to the mean() function:

gss %>%
group_by(sex) %>%
  summarise(xbar = mean(educ, na.rm = TRUE))

[aephiday25]

If the Education variable is categorical but in form of numerical data, it's meaningless in statistics to calculate the mean of categorical data.

[fcas80]

Use the str function to determine whether educ is a factor or a number:

str(gss$educ)

If it is looks like a number but nevertheless is a factor, then convert it to numeric in order to calculate its mean:

gss$educ <- as.numeric(gss$educ)

[SL_Mabaso]

Thank you this way was much easier for me to use.