Detect potentially difficult questions

Filters summarise_questions() to questions whose mean percent is below a threshold and that have enough observed activity.

Usage

detect_difficult_questions(
  con,
  tutorial_id,
  rule = c("last", "best", "first"),
  include_unregistered = TRUE,
  group_id = NULL,
  max_mean_percent = 60,
  min_students = 1,
  min_attempts = 1
)

Arguments

con

A DBI connection.

tutorial_id

Tutorial identifier. Required.

rule

Scoring rule used to select one attempt per student and question.

include_unregistered

If TRUE, include attempted questions not registered with register_questions().

group_id

Optional registered student group identifier. If supplied, only attempts from students in that group are included.

max_mean_percent

Maximum mean percent used to flag a question. Defaults to 60.

min_students

Minimum number of students with a selected attempt. Defaults to 1.

min_attempts

Minimum number of raw attempts. Defaults to 1.

Value

A tibble of flagged question summaries, sorted from lowest mean percent to highest.

Examples

db_path <- tempfile(fileext = ".sqlite")
con <- init_tracking_db(db_path, overwrite = TRUE)
track_attempt(con, "student_001", "module_01", "q1", "wrong", score = 0, max_score = 1)
detect_difficult_questions(con, "module_01")
#> # A tibble: 1 × 14
#>   tutorial_id question_id question_label question_type max_score
#>   <chr>       <chr>       <chr>          <chr>             <dbl>
#> 1 module_01   q1          q1             NA                    1
#> # ℹ 9 more variables: n_possible_students <int>, n_students <int>,
#> #   n_attempts <int>, n_answered <int>, n_full_credit <int>, mean_score <dbl>,
#> #   mean_percent <dbl>, full_credit_rate <dbl>, mean_attempts_per_student <dbl>
DBI::dbDisconnect(con)