Add Marginal Observations

Given a data frame and the names of variable, doubles the data frame for each variable with a new category "All" by default, or by the value of label. A new variable .marginal. is added to the resulting data frame, with value "" if the observation is an original one, and with value equal to the names of the variable being marginalized (separated by commas) otherwise. If there is another stratification variable besides the one in ..., and that variable is nested inside the variable in ..., specify nested=variable name to have the value of that variable set fo label whenever marginal observations are created for .... See the state-city example below.

Usage

addMarginal(data, ..., label = "All", margloc=c('last', 'first'), nested)

Arguments

data

a data frame

...

a list of names of variables to marginalize

label

category name for added marginal observations

margloc

location for marginal category within factor variable specifying categories. Set to "first" to override the default - to put a category with value label as the first category.

nested

a single unquoted variable name if used

Examples

d <- expand.grid(sex=c('female', 'male'), country=c('US', 'Romania'),
                 reps=1:2)
addMarginal(d, sex, country)
#>       sex country reps  .marginal.
#> 1  female      US    1            
#> 2    male      US    1            
#> 3  female Romania    1            
#> 4    male Romania    1            
#> 5  female      US    2            
#> 6    male      US    2            
#> 7  female Romania    2            
#> 8    male Romania    2            
#> 9     All      US    1         sex
#> 10    All      US    1         sex
#> 11    All Romania    1         sex
#> 12    All Romania    1         sex
#> 13    All      US    2         sex
#> 14    All      US    2         sex
#> 15    All Romania    2         sex
#> 16    All Romania    2         sex
#> 17 female     All    1     country
#> 18   male     All    1     country
#> 19 female     All    1     country
#> 20   male     All    1     country
#> 21 female     All    2     country
#> 22   male     All    2     country
#> 23 female     All    2     country
#> 24   male     All    2     country
#> 25    All     All    1 sex,country
#> 26    All     All    1 sex,country
#> 27    All     All    1 sex,country
#> 28    All     All    1 sex,country
#> 29    All     All    2 sex,country
#> 30    All     All    2 sex,country
#> 31    All     All    2 sex,country
#> 32    All     All    2 sex,country

# Example of nested variables
d <- data.frame(state=c('AL', 'AL', 'GA', 'GA', 'GA'),
                city=c('Mobile', 'Montgomery', 'Valdosto',
                       'Augusta', 'Atlanta'),
                x=1:5, stringsAsFactors=TRUE)
addMarginal(d, state, nested=city) # cite set to 'All' when state is
#>    state       city x .marginal.
#> 1     AL     Mobile 1           
#> 2     AL Montgomery 2           
#> 3     GA   Valdosto 3           
#> 4     GA    Augusta 4           
#> 5     GA    Atlanta 5           
#> 6    All        All 1      state
#> 7    All        All 2      state
#> 8    All        All 3      state
#> 9    All        All 4      state
#> 10   All        All 5      state