Skip to contents

Grading Exams

Alejandro Gonzalez Recuenco

2024-01-23

Source: vignettes/GradingExams.Rmd
GradingExams.Rmd

This vignette describes how to grade an exam using the script found on the exec/ folder. You can find where this script is found by typing on an R terminal.

    system.file("exec", package = "TexExamRandomizer")
## [1] "/home/runner/work/_temp/Library/TexExamRandomizer/exec"

Format of the student’s answers

When you collect the responses from the student, you should follow the following prescriptions:

  1. The actual answers to each question should be an integer number, (1 being the first choice). They should be written in the order that they are presented in the student’s exam. The columns should be called “Question 1”, “Question 2”, etc. (Or “Q 1”, “Q2”, etc).
    • If a certain question is left unanswered, give it the value 0 in the table.
  2. If you want to give extra points, add a column called “Extra Points” with the added points.
  3. You require a column called “Version”, where the version number for each exam is written.
  4. All other columns will be placed on the output “_Graded.csv” table.
Example responses format
Class Exam.Version Nickname Roll.Number Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Q15 ExtraPoints
M601 21 Titar 1 3 4 4 1 1 1 1 3 4 1 3 4 1 2 4 9
M601 1 Suwannapoe 3 2 3 1 2 3 3 2 4 4 2 3 3 2 2 3 8
M601 5 Kan 3 3 3 4 2 1 3 4 2 1 1 3 4 2 3 3 8
M601 16 Mei 16 1 1 2 4 1 2 3 2 1 3 1 3 2 3 1 5
M601 17 Offside 6 2 4 4 1 3 2 1 2 3 3 2 1 1 2 3 1

Format of the answer sheet

When you created the exams by using the examrandomizer script, it will generate a fullanswersheet.csv file. When grading, it assumes that the “correctTag” and the “wrongTag” are respectively “choice” and “CorrectChoice”.

If that is not the case in your structure, simply change the names to “choice” and “CorrectChoice” of those columns.

Answer sheet, the last two rows must be named “choice” and “CorrectChoice”
Version index questions_original question_original questions question choices_original X.choice.CorrectChoice._original choices choice CorrectChoice
0 1 1 1 1 1 1 1 1 1 NA
0 2 1 1 1 1 1 2 1 2 NA
0 3 1 1 1 1 1 3 1 3 NA
0 4 1 1 1 1 1 4 1 NA 4
0 5 1 2 1 2 1 1 1 1 NA

(But if you really want to change that assumption, you can access the script and edit ASHEET_COLCORRECT and ASHEET_COLINCORRECT)

What if I wrote some questions wrong and I noticed too late? I already printed the exams

If you realize one question is incorrectly written, but it is too late to rewrite the exam, find the original answer sheet in the fullanswersheet.csv (The rows with version number 0). Then, remove the lines from the answer sheet that refer to the question you want to ignore. (Keep a backup of the answer sheet, just in case).

When the students answer the exam. Tell them to write any answer in that question, it will be ignored by the program.

Grading the exam

You can specify directly the responses from the students (--resp) and the answer sheet (--answer).

gradeexamrandomizer --resp <student responses csv> --answer <fullanswersheet csv>

If you have both of those files on the same directory and you make sure their name is somehow similar to “responses*.csv” and “answer*.csv”, you can write the shorthand version.

gradeexamrandomizer --dir <dirname>

The output will be two csv files in the same directory where the students’ responses file is found, called *_Graded.csv and *_Stats.csv.

Output graded file
Class Exam.Version Nickname Roll.Number addedPoints addedAllPoints maxGrade Grade Grade_Total_Exam
M601 21 Titar 1 0 0 14 14 100.00000
M601 1 Suwannapoe 3 0 0 14 10 71.42857
M601 5 Kan 3 0 0 15 5 33.33333
M601 16 Mei 16 0 0 15 8 53.33333
M601 17 Offside 6 0 0 14 6 42.85714
Output stats (It simply counts how many times each choice was chosen by every student)
Version index questions_original question_original questions question choices_original X.choice.CorrectChoice._original choices choice CorrectChoice ExamAnswerCount
0 1 1 1 1 1 1 1 1 1 NA 1
0 2 1 1 1 1 1 2 1 2 NA 1
0 3 1 1 1 1 1 3 1 3 NA 5
0 4 1 1 1 1 1 4 1 NA 4 18
0 5 1 2 1 2 1 1 1 1 NA 1

How is the grade calculated

In the output *_Graded.csv table, as you can see in the example above, there will be 5 rows added to the output:

  • addedPoints: Points added to each student, by adding a column in the responses called “ExtraPoints”

  • addedAllPoints: Points added to all students. This points are added differently, like if the exam had addedAllPoints extra questions, and all students got them correctly.

    (Which right now will always be zero with the current script).

  • maxGrade: Maximum number of answers in that exam. If you had removed some questions that you want to remove, but that question is not found on all versions, then some exams will have a greater than

  • Grade: The number of correct answers on the student exam. (If there are more than one correct answer on each question, ti won’t be able to detec all of them, only whether the student wrote a correct one or not)

  • Grade_Total_Exam: The total grade, after scaling. It assumes the maximum grade of the exam is a 100%. For generailty we will denote it MAX. (To change it, use the option --max <integer>) \[Grade_{Total Exam} = \frac{Grade + addedAllPoints}{maxGrade + addedAllPoints} \cdot MAX + addedPoints\]