Blog
AI multiple choice grading covers the letters on a paper test that also asks for short answers. Students circle a letter or write it on the line, and on the same page they write a word, a date, a formula, or a short explanation. The letter is a list you can write in advance. The short answer is that list, or a response you score for quality.
Graider is the AI red pen for paper grading, a product of DM Liquidity LLC. You upload the test and your answer key or rubric, photograph the stack of student papers, and Graider marks each question against that key. Each question gets one score and one comment. You review and override any mark, toggle feedback, keep the pen, and hand the work back as a PDF or an email. Teachers who grade multiple choice tests with AI on ordinary paper follow that same path in AI test grading.
Your key. Their papers. You review every mark. You keep the pen.
An answer key is enough when you can list every response that earns points before you photograph the stack. Multiple choice sits in that group. So does a short factual answer with a closed list: a name, a date, a formula, a vocabulary word, or a definition you have already decided to accept.
For a multiple choice item, the key lists the letter that earns the points, whether students circle B or write B. Option words you also accept are another line with the same points. A second full-credit letter is its own line. A distractor that earns partial credit is a line with those points. The matching line is the one score for that question.
A blank or a response absent from the key is one you score yourself at review. When a new wording counts, override the score on that paper, and list that wording on your key for the next test.
Write the key in the order of the test. Give each scored response its own line: the question, the response, and the points.
Question 1 is B for 1 point. If "mitochondria" also earns that point, add it as a second line. List B and b when you accept both. When students choose two letters, write the full-credit pair, such as A and C for 2 points, and A alone for 1 point when that is your partial. The question still receives one score.
Put the wording you want beside the answer on the key, such as "B. The mitochondria make ATP." Graider returns one comment for the question. Edit it at review so it sounds like you.
A letter that refers to a diagram, a graph, or a drawing still belongs on the key. Graider does not read or score diagrams, graphs, or drawings. When the item says "sketch the graph" or "draw the setup," set those points yourself inside the one score for that question.
Short answer grading with AI still starts with a key when the acceptable responses are a list. Put the full-credit wording on the key, then every equivalent you count, then the partial versions with their points.
A vocabulary item can list "evaporation" for 2 points, and "water turns to vapor" for 2 points when that sentence counts as the same answer in your class. A math result can list "12 cm" for 2 points, and "12" for 1 point when a missing unit costs a point. Those points are the one score for the question.
List "H2O" and "water" both when they earn the same mark. When you accept a phrasing that is new, override the score on that paper at review, and list that wording on your key for the next test. How to keep that list stable is in short answer grading with AI.
A rubric carries the points when many wordings can be right. An explanation, a claim with evidence, or a method belongs on a rubric. Say what a full response includes, what a middle score includes, and the points for each.
On "Explain why the plant grew taller in the window," full credit can be 4 points: the response names the light and says how it changes growth. Put an example of a 2 point response beside the rubric on the key: the response names the light and leaves the change unstated. The rubric and those examples feed the one score and the one comment for the question. You review that score and edit that comment.
A mixed paper uses both documents. Letters and listed facts stay on the answer key. The explanation stays on the rubric. Upload both with the test. If the choices are 20 points and the explanation is 10, the paper is 30, and each part should add to that total. How a fixed key and a rubric divide one paper is in answer key grading vs rubric.
Photograph every page in even light, with the full sheet in the frame. You confirm the roster match, then review before you hand anything back.
Review shows one score and one comment per question. Keep a score when it matches your key or rubric. Override it when you disagree, and edit the comment. Toggle feedback on when the comment should be on the PDF or email, and off when you will teach the item with the class.
The only flag is the Low confidence badge on an answer that Graider found hard to read. Read the student's answer as text, compare it with the paper and your key, and override when you disagree. If the writing is on the sheet and unmatched to a question, set the points yourself.
Before you finish review, check:
You keep the pen on every override.
Hand each paper back as a PDF or an email after you finish review. The score shows the points you kept for each question. When feedback is on, the comment you kept sits with that question.
AI multiple choice grading marks paper tests where students circle a letter or write it on the line. Graider marks each question against your answer key and returns one score and one comment. You review and override any mark, toggle feedback, and hand the work back as a PDF or an email.
An answer key is enough when you can list every response that earns points: the letter, any option text you also accept, and any partial-credit letter with its points. A short factual answer with a closed list of wordings belongs on that same key.
List accepted variants on the key when you can write each full-credit wording and each partial-credit wording in advance. Use a rubric when you score the quality of an explanation. The rubric still feeds one score and one comment for that question.
Open every Low confidence badge. It is the only flag, and it sits on a hard-to-read answer. Compare the read with the paper and your key, override any score you disagree with, and edit the comment. Set the points yourself on a diagram, a graph, or a drawing.
Check every Low confidence badge, override any mark you disagree with, and toggle feedback. You keep the pen.