The data should be 1 row per respondent and 1 column per item, with an
optional additional column to store respondent identifiers. Each value of the
data should be a 0 or 1 to indicate the response to the item by the given
respondent. clean_data() calls check_data() to verify the expected
structure, and then performs additional data manipulation to provide standard
conventions. See details for additional information.
Usage
check_data(
x,
identifier = NULL,
missing = NA,
arg = rlang::caller_arg(x),
call = rlang::caller_env()
)
clean_data(
x,
identifier = NULL,
missing = NA,
cleaned_qmatrix,
arg_qmatrix = rlang::caller_arg(cleaned_qmatrix),
arg = rlang::caller_arg(x),
call = rlang::caller_env()
)Arguments
- x
The provided data to check.
- identifier
The provided respondent identifier, as a character string. If no respondent identifier is present, the value should be
NULL(the default).- missing
A expression specifying how missing data in
xis coded (e.g.,NA,".",-99). The default isNA.- arg
The name of the argument.
- call
The call stack.
- cleaned_qmatrix
A cleaned Q-matrix, from
clean_qmatrix().- arg_qmatrix
A character string with the name of the argument used to provide the Q-matrix.
Value
check_data returns the original data (if the checks pass) as a
tibble, with missing data (i.e., missing)
replaced with NA.
clean_data returns a list with five elements:
clean_data: The cleaned dataitem_identifier: The real name of the item identifieritem_names: The real names of the itemsrespondent_identifier: The real name of the respondent identifierrespondent_names: The real names of the respondents
Details
In many instances, it's important to have standard conventions for a data
object so that we know what to expect (e.g., respondent and item identifiers,
data types). clean_data() provides this standardization. Cleaned data is
returned in long format, with one row per response. Respondent and item
columns are encoded as factors, and responses are coerced to integer values.
To ensure downstream functions are able to identify the original
(pre-cleaned) values, clean_data() returns a list that includes the cleaned
data, as well as metadata that includes look-ups from the original to cleaned
values.
Examples
example_data <- tibble::tibble(person = 1:10,
item1 = sample(0:1, 10, replace = TRUE),
item2 = sample(0:1, 10, replace = TRUE),
item3 = sample(0:1, 10, replace = TRUE))
check_data(example_data, identifier = "person")
#> # A tibble: 10 × 4
#> person item1 item2 item3
#> <int> <int> <int> <int>
#> 1 1 0 1 0
#> 2 2 0 0 0
#> 3 3 1 0 1
#> 4 4 0 0 0
#> 5 5 0 0 0
#> 6 6 1 1 0
#> 7 7 0 0 1
#> 8 8 0 0 0
#> 9 9 0 1 1
#> 10 10 1 1 1
example_qmatrix <- tibble::tibble(item = paste0("item", 1:3),
att_1 = c(0, 0, 1),
att_2 = c(1, 1, 1))
example_data <- tibble::tibble(person = 1:10,
item1 = sample(0:1, 10, replace = TRUE),
item2 = sample(0:1, 10, replace = TRUE),
item3 = sample(0:1, 10, replace = TRUE))
qmatrix <- clean_qmatrix(example_qmatrix, identifier = "item")
clean_data(example_data, identifier = "person",
cleaned_qmatrix = qmatrix)
#> $clean_data
#> # A tibble: 30 × 3
#> resp_id item_id score
#> <fct> <fct> <int>
#> 1 1 item1 0
#> 2 1 item2 1
#> 3 1 item3 0
#> 4 2 item1 0
#> 5 2 item2 0
#> 6 2 item3 1
#> 7 3 item1 1
#> 8 3 item2 1
#> 9 3 item3 1
#> 10 4 item1 1
#> # ℹ 20 more rows
#>
#> $item_identifier
#> [1] "item"
#>
#> $item_names
#> item1 item2 item3
#> 1 2 3
#>
#> $respondent_identifier
#> [1] "person"
#>
#> $respondent_names
#> 1 2 3 4 5 6 7 8 9 10
#> 1 2 3 4 5 6 7 8 9 10
#>