Creating variables

Basic command to create a variable

To create a new variable called new_variable inside a dataframe called df:

df$new_variable <- ...

Replace the ... with the formula you want to use to define the new variable.

If you try to create a new variable using the name of an existing variable, you will overwrite the existing variable with the new variable. Since this causes data loss, you should almost never do this!

Creating a boolean variable

One of the most common things you’ll do in this class is to create boolean variables that indicate whether some condition in the data is true.

For example, in IPUMS, the variable EMPSTAT is equal to 1 if the person is employed.

To create a boolean variable called EMPLOYED that is TRUE if the person is employed and FALSE otherwise, you can do this:

df$EMPLOYED <- df$EMPSTAT==1

See here for a more detailed explanation of logical operators.

Creating a variable with mathematical expressions

Sometimes, we need to take the log of a variable. To do this, we can create a new variable that is equal to the log of another variable:

df$LOG_INCWAGE <- log(df$INCWAGE)

We could have chosen any name for the new variable. df$foo <- log(df$INCWAGE) would have done the same thing, just naming the variable foo instead of LOG_INCWAGE.

We can create variables using arbitrary mathematical operations. For example, suppose the year of the survey is contained in the variable YEAR and the age of the person is contained in the variable AGE. The person’s birth year is equal to YEAR - AGE, so we can create a variable called BIRTH_YEAR like this:

df$BIRTH_YEAR <- df$YEAR - df$AGE

Operations work row-by-row

When creating a variable, the operations are performed row-by-row. For example, df$BIRTH_YEAR <- df$YEAR - df$AGE calculates YEAR - AGE for each row and puts it in the corresponding row of BIRTH_YEAR.

Example:

Suppose the dataframe called df contains the following data:

ID EMPSTAT INCWAGE YEAR AGE
1 1 100000 2019 30
2 3 0 2019 40
3 2 20000 2019 25
4 1 80000 2019 50

If you run:

df$EMPLOYED <- df$EMPSTAT==1
df$LOG_INCWAGE <- log(df$INCWAGE)
df$BIRTH_YEAR <- df$YEAR - df$AGE

Then after running these commands, df will contain:

ID EMPSTAT INCWAGE YEAR AGE EMPLOYED LOG_INCWAGE BIRTH_YEAR
1 1 100000 2019 30 TRUE 11.513 1989
2 3 0 2019 40 FALSE NA 1979
3 2 20000 2019 25 FALSE 9.903 1994
4 1 80000 2019 50 TRUE 11.290 1969