Compute the Average Rank of All Items

This function calculates the average rank for the data frame that contains ranking data. It can be used for both long- and wide-type data frames.

Usage

avg_rank(
  x,
  rankings = NULL,
  items = NULL,
  long = FALSE,
  raw = TRUE,
  weight = NULL,
  round = NULL
)

Arguments

x

A data frame that contains rankings of items.

rankings

The name of the column that contains the rankings. Defaults to NULL, which means that the function will look for a data frame with two columns, "item" and "rank". For wide data, this can also be a character vector of separate ranking columns such as c("apple", "orange", "banana").

items

The name of the column that contains the items' names, or, in case of a wide file, the item names in the reference choice set. Defaults to NULL.

long

The type of the data frame. Defaults to `FALSE`. It `TRUE`, which means that the data frame is in the long format, it is presumed to be generated by rank_longer(). If the data frame is in the wide format, it should be set to `FALSE`. For wide data, rankings can be supplied either as a single encoded ranking string such as "123" or as separate ranking columns.

raw

If TRUE, the function will return the raw average rank. If FALSE, the function will return the average rank after correcting based on the IPW estimator. Defaults to TRUE.

weight

Optional weight specification. This can be the name of a weight column in x or a numeric vector with one weight per row. Defaults to NULL.

round

The number of decimal places to round the output to. Defaults to NULL.

Value

A data frame with the average rank of each item in the reference choice set.

Examples

x <- data.frame(
  id = c("Bernie", "Yuki", "Silvia"),
  rank = c("123", "321", "213")
)
avg_rank(x, "rank")
#>   item          qoi     mean        se     lower    upper   method
#> 1  1st Average Rank 2.000000 0.5773503 0.8683935 3.131607 Raw Data
#> 2  2nd Average Rank 1.666667 0.3333333 1.0133333 2.320000 Raw Data
#> 3  3rd Average Rank 2.333333 0.6666667 1.0266667 3.640000 Raw Data
avg_rank(x, "rank", items = c("Money", "Power", "Respect"))
#>      item          qoi     mean        se     lower    upper   method
#> 1   Money Average Rank 2.000000 0.5773503 0.8683935 3.131607 Raw Data
#> 2   Power Average Rank 1.666667 0.3333333 1.0133333 2.320000 Raw Data
#> 3 Respect Average Rank 2.333333 0.6666667 1.0266667 3.640000 Raw Data

y <- data.frame(rank = c("123", "321", "213"))
avg_rank(y, "rank")
#>   item          qoi     mean        se     lower    upper   method
#> 1  1st Average Rank 2.000000 0.5773503 0.8683935 3.131607 Raw Data
#> 2  2nd Average Rank 1.666667 0.3333333 1.0133333 2.320000 Raw Data
#> 3  3rd Average Rank 2.333333 0.6666667 1.0266667 3.640000 Raw Data

x_sep <- data.frame(
  apple = c(2, 1, 3),
  orange = c(1, 3, 2),
  banana = c(3, 2, 1)
)
avg_rank(x_sep, rankings = c("apple", "orange", "banana"))
#>     item          qoi mean        se     lower    upper   method
#> 1  apple Average Rank    2 0.5773503 0.8683935 3.131607 Raw Data
#> 2 orange Average Rank    2 0.5773503 0.8683935 3.131607 Raw Data
#> 3 banana Average Rank    2 0.5773503 0.8683935 3.131607 Raw Data

x_weighted <- data.frame(
  rank = c("12", "21"),
  survey_weight = c(1, 3)
)
avg_rank(x_weighted, "rank", weight = "survey_weight")
#>   item          qoi mean        se     lower    upper   method
#> 1  1st Average Rank 1.75 0.4330127 -3.751948 7.251948 Raw Data
#> 2  2nd Average Rank 1.25 0.4330127 -4.251948 6.751948 Raw Data
z <- rank_longer(
  y,
  cols = "rank",
  reference = c("Money", "Power", "Respect")
)
#> No ID column specified. Using row number.
#> One column selected. Parsing encoded ranking values.
avg_rank(z, "ranking", items = "item_name", long = TRUE)
#>   item_name          qoi     mean        se     lower    upper   method
#> 1     Money Average Rank 2.000000 0.5773503 0.8683935 3.131607 Raw Data
#> 2     Power Average Rank 1.666667 0.3333333 1.0133333 2.320000 Raw Data
#> 3   Respect Average Rank 2.333333 0.6666667 1.0266667 3.640000 Raw Data

## Example output from item_to_rank
x <- data.frame(
  item = c("a", "b", "c", "a", "b", "c", "a", "b", "c"),
  rank = c(3L, 1L, 2L, 1L, 2L, 3L, 3L, 2L, 1L)
)
avg_rank(x, long = TRUE)
#>   item          qoi     mean        se     lower    upper   method
#> 1    a Average Rank 2.333333 0.6666667 1.0266667 3.640000 Raw Data
#> 2    b Average Rank 1.666667 0.3333333 1.0133333 2.320000 Raw Data
#> 3    c Average Rank 2.000000 0.5773503 0.8683935 3.131607 Raw Data