Function to combine several items (e.g., by taking their mean).

combitems(items, data, grep=FALSE, fun=mean, na.rm=TRUE, min.k, verbose=TRUE)

Arguments

items

either a character vector (giving the names of the variables to be combined) or a numeric vector (giving the position of the columns in the data frame corresponding to the items).

data

data frame that contains the items to be combined.

grep

logical indicating whether item names should be matched using grep (default is FALSE).

fun

the function to be applied to the items (default is mean).

na.rm

logical indicating whether missing values should be removed before applying the function (default is TRUE).

min.k

optional integer specifying the minimum number of items that must be available (i.e., not missing). If not specified, no minimum is required.

verbose

logical indicating whether information about the items that were combined should be reported (default is TRUE).

Details

The function can be used to combine several items (usually by taking their mean or their sum) into a new variable (the default is to take the mean).

The items to be combined are specified via the items argument, either as a character vector with the variable names or as a numeric vector giving the item position in the data frame. One can also just specify substrings for the item names, which are matched (if possible) against the variable names in the data frame. See ‘Examples’.

The function to be applied to the items is specified via the fun argument and must be a function that returns a single number.

At times, one may require that a certain number of items must be available (i.e., not missing) before computing the function. This can be specified via the min.k argument.

Value

The function returns a vector (that will typically be added back to the data frame as a new variable).

Author

Wolfgang Viechtbauer wvb@wvbauer.com

Examples

# illustrative dataset dat <- structure(list(id=1:10, mood_cheerf = c(5, 6, 6, 6, 7, 6, 7, 7, 6, 7), mood_relaxed = c(6, 6, 6, 7, 7, NA, 5, 5, 5, 6), mood_satisfi = c(6, 6, 5, NA, 7, NA, 6, 6, 5, 6)), row.names = 1:10, class = "data.frame") dat
#> id mood_cheerf mood_relaxed mood_satisfi #> 1 1 5 6 6 #> 2 2 6 6 6 #> 3 3 6 6 5 #> 4 4 6 7 NA #> 5 5 7 7 7 #> 6 6 6 NA NA #> 7 7 7 5 6 #> 8 8 7 5 6 #> 9 9 6 5 5 #> 10 10 7 6 6
# compute the average of the three mood items and add # this as a new variable (called 'pa') to the dataset dat$pa <- combitems(c("mood_cheerf", "mood_relaxed", "mood_satisfi"), data=dat)
#> Computed 'mean' for these items: mood_cheerf, mood_relaxed, mood_satisfi.
dat
#> id mood_cheerf mood_relaxed mood_satisfi pa #> 1 1 5 6 6 5.666667 #> 2 2 6 6 6 6.000000 #> 3 3 6 6 5 5.666667 #> 4 4 6 7 NA 6.500000 #> 5 5 7 7 7 7.000000 #> 6 6 6 NA NA 6.000000 #> 7 7 7 5 6 6.000000 #> 8 8 7 5 6 6.000000 #> 9 9 6 5 5 5.333333 #> 10 10 7 6 6 6.333333
# can also grep for the item names dat$pa <- combitems(c("cheerf", "relaxed", "satisfi"), data=dat, grep=TRUE)
#> Computed 'mean' for these items: mood_cheerf, mood_relaxed, mood_satisfi.
dat
#> id mood_cheerf mood_relaxed mood_satisfi pa #> 1 1 5 6 6 5.666667 #> 2 2 6 6 6 6.000000 #> 3 3 6 6 5 5.666667 #> 4 4 6 7 NA 6.500000 #> 5 5 7 7 7 7.000000 #> 6 6 6 NA NA 6.000000 #> 7 7 7 5 6 6.000000 #> 8 8 7 5 6 6.000000 #> 9 9 6 5 5 5.333333 #> 10 10 7 6 6 6.333333
# even shorter dat$pa <- combitems("mood", data=dat, grep=TRUE)
#> Computed 'mean' for these items: mood_cheerf, mood_relaxed, mood_satisfi.
dat
#> id mood_cheerf mood_relaxed mood_satisfi pa #> 1 1 5 6 6 5.666667 #> 2 2 6 6 6 6.000000 #> 3 3 6 6 5 5.666667 #> 4 4 6 7 NA 6.500000 #> 5 5 7 7 7 7.000000 #> 6 6 6 NA NA 6.000000 #> 7 7 7 5 6 6.000000 #> 8 8 7 5 6 6.000000 #> 9 9 6 5 5 5.333333 #> 10 10 7 6 6 6.333333
# can also just specify the position of the items in the dataset dat$pa <- combitems(2:4, data=dat)
#> Computed 'mean' for these items: mood_cheerf, mood_relaxed, mood_satisfi.
dat
#> id mood_cheerf mood_relaxed mood_satisfi pa #> 1 1 5 6 6 5.666667 #> 2 2 6 6 6 6.000000 #> 3 3 6 6 5 5.666667 #> 4 4 6 7 NA 6.500000 #> 5 5 7 7 7 7.000000 #> 6 6 6 NA NA 6.000000 #> 7 7 7 5 6 6.000000 #> 8 8 7 5 6 6.000000 #> 9 9 6 5 5 5.333333 #> 10 10 7 6 6 6.333333
# can specify the minimum number of required non-missing items # note: the mean for id=6 is NA because only one item is non-missing dat$pa <- combitems("mood", data=dat, grep=TRUE, min.k=2)
#> Computed 'mean' for these items: mood_cheerf, mood_relaxed, mood_satisfi.
dat
#> id mood_cheerf mood_relaxed mood_satisfi pa #> 1 1 5 6 6 5.666667 #> 2 2 6 6 6 6.000000 #> 3 3 6 6 5 5.666667 #> 4 4 6 7 NA 6.500000 #> 5 5 7 7 7 7.000000 #> 6 6 6 NA NA NA #> 7 7 7 5 6 6.000000 #> 8 8 7 5 6 6.000000 #> 9 9 6 5 5 5.333333 #> 10 10 7 6 6 6.333333