non_functional_distractors_pos_cor: Non-Functional Distractors (Positive Correlations)

In melissavanbussel/Analyze-Scantron: Analyze your multiple choice test

Description

Finds the non-functional distractors on the test which had a positive correlation with student scores.

Usage

non_functional_distractors_pos_cor(exam, numchoicesperitem = NULL,
  check_format = TRUE)

Arguments

exam	A data.frame object containing a test that follows the formatting of a typical Scantron file.
numchoicesperitem	A vector which has the same length as the number of questions on the test, where each element in the vector corresponds to the number of choices there were for each of the questions on the test. For example, if your test had 35 questions which each had 5 options, numchoicesperitem = rep(5, 35).
check_format	a logical value (default = TRUE) indicating whether the "exam" object should be tested for correct formatting.

Value

Returns a list object which is the same length as the number of questions on the test. Each element in the list is a vector corresponding to one of the questions on the test, where this vector contains the non-functional distractors which had positive correlations with student scores for that question.

Note

The argument "exam" must be in the format of a typical Scantron results file. Column 1 should correspond to ID (e.g., student number); column 2 should correspond to DEPT (e.g., MATH); column 3 should correspond to COURSE CODE (e.g., 1051); the remaining columns should each correspond to one of the questions on the test. The header of the data frame should contain the column names, and row 1 of the data frame should contain the answer key for the test. For example, if you had an exam with 25 students and 40 questions, the data.frame object should have 26 rows and 43 columns.

The "check_format" argument defaults to null. If this is left as null, the function will call the num_choices_per_item function, which will do its best to guess the number of options for each question. This is done by looking at the student answers and finding the "largest" answer for each question. For example, if at least one student answered "E", but no students answered "F", the function would guess that there were 5 options for that question.

melissavanbussel/Analyze-Scantron documentation built on May 10, 2019, 1:19 a.m.