Levels of Measurement of Survey Items

The measurement level of a "item" object, which is one of "nominal", "ordinal", "interval", "ratio", determines what happens to it, if it or the data.set containing it is coerced into a data.frame. If the level of measurement level is "nominal", the it will be converted into an (unordered) factor, if the level of measurement is "ordinal", the item will be converted into an ordered vector. If the measurement is "interval" or "ratio", the item will be converted into a numerical vector.

Usage

# S4 method for item
measurement(x)
# S4 method for item
measurement(x) <- value
# S4 method for data.set
measurement(x)
# S4 method for data.set
measurement(x) <- value
is.nominal(x)
is.ordinal(x)
is.interval(x)
is.ratio(x)
as.nominal(x)
as.ordinal(x)
as.interval(x)
as.ratio(x)
set_measurement(x,...)

Arguments

x

an object, usually of class "item".

value

for the item method, a character string; either "nominal", "ordinal", "interval", or "ratio"; for the data.set method, a list of character vectors with variable names, where the names of the list corresponds to a measurement level and and the list elements indicates the variables to which the measurement levels are assigned.

...

vectors of variable names, either symbols or character strings, tagged with the intended measurement level.

Value

The item method of measurement(x) returns a character string, the data.set method returns a named character vector, where the name of each element is a variable name and each.

as.nominal, as.ordinal, as.interval, as.ratio

return an item with the requested level of measurement setting.

is.nominal, is.ordinal, is.interval, is.ratio

return a logical value.

References

Stevens, Stanley S. 1946. "On the theory of scales of measurement." Science 103: 677-680.

See also

data.set, item

Examples

vote <- sample(c(1,2,3,8,9),size=30,replace=TRUE)
labels(vote) <- c(Conservatives         =  1,
                  Labour                =  2,
                  "Liberal Democrats"   =  3,
                  "Don't know"          =  8,
                  "Answer refused"      =  9
                  )
missing.values(vote) <- c(8,9)
as.data.frame(vote)[[1]]
#>  [1] <NA>              <NA>              <NA>              Labour           
#>  [5] <NA>              <NA>              Liberal Democrats Liberal Democrats
#>  [9] <NA>              Labour            Liberal Democrats Labour           
#> [13] <NA>              Conservatives     Liberal Democrats <NA>             
#> [17] <NA>              <NA>              Labour            <NA>             
#> [21] Labour            <NA>              Labour            Liberal Democrats
#> [25] Labour            Liberal Democrats Liberal Democrats <NA>             
#> [29] <NA>              <NA>             
#> Levels: Conservatives Labour Liberal Democrats
measurement(vote) <- "interval"
as.data.frame(vote)[[1]]
#>  [1] NA NA NA  2 NA NA  3  3 NA  2  3  2 NA  1  3 NA NA NA  2 NA  2 NA  2  3  2
#> [26]  3  3 NA NA NA
vote <- as.nominal(vote)
as.data.frame(vote)[[1]]
#>  [1] <NA>              <NA>              <NA>              Labour           
#>  [5] <NA>              <NA>              Liberal Democrats Liberal Democrats
#>  [9] <NA>              Labour            Liberal Democrats Labour           
#> [13] <NA>              Conservatives     Liberal Democrats <NA>             
#> [17] <NA>              <NA>              Labour            <NA>             
#> [21] Labour            <NA>              Labour            Liberal Democrats
#> [25] Labour            Liberal Democrats Liberal Democrats <NA>             
#> [29] <NA>              <NA>             
#> Levels: Conservatives Labour Liberal Democrats
group <- sample(c(1,2),size=30,replace=TRUE)
labels(group) <- c(A=1,B=2)
DataS <- data.set(group,vote)
measurement(DataS)
#> NULL
measurement(DataS) <- list(interval=c("group","vote"))
head(as.data.frame(DataS))
#>   group vote
#> 1     2   NA
#> 2     1   NA
#> 3     2   NA
#> 4     1    2
#> 5     2   NA
#> 6     1   NA
DataS <- set_measurement(DataS,
                         nominal=c(group,vote))
head(as.data.frame(DataS))
#>   group   vote
#> 1     B   <NA>
#> 2     A   <NA>
#> 3     B   <NA>
#> 4     A Labour
#> 5     B   <NA>
#> 6     A   <NA>