Blog
Fill in the blank grading covers paper tests where a student writes a word, a name, a date, a number, a formula, or a short phrase on a line. A vocabulary quiz, a grammar cloze, a science recall sheet, and a unit check with a row of blanks all sit in this group. The points live on the answer key you already use. The papers are photographs of what students wrote on those lines.
Graider is the AI red pen for paper grading, a product of DM Liquidity LLC. You upload the test and your answer key, 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 fill in the blank 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.
Write the key in the order of the test. Give each scored response its own line: the question, the response, and the points.
Question 4 is "photosynthesis" for 1 point. If you also accept "plants use light to make food" for that same point, add it as a second line. A shorter response you count for less, such as "plants make food," is its own line with the points you assign. The line that matches the paper is the one score for that question.
Put the wording you want beside the answer on the key, such as "photosynthesis: plants use light to make food." Graider returns one comment for the question. Edit that comment at review so it sounds like you.
A blank left empty is a response you score yourself at review. A word absent from the key is the same kind of case. When a new wording counts, override the score on that paper, and list that wording on your key for the next test.
Some blanks sit beside a figure. A label that is a word you can list still belongs on the key, such as "nucleus" when the word itself is what you score. Graider does not read or score diagrams, graphs, or drawings. When the item says "draw the cell" or "sketch the graph," set those points yourself inside the one score for that question.
Accepted answers are the lines you write on the key. They include the response you expect and every variant you have decided to count.
A capital item can list "Paris" for 1 point. If you also accept "paris" and "Paris, France," each of those is another line with the same points. A date item can list "1776" for 1 point, and "July 4, 1776" for 1 point when the longer form counts as the same answer. A formula can list "H2O" for 2 points and "water" for 2 points when both earn the mark.
Spelling variants belong on the key in the same way. On a vocabulary blank, list "receive" for 2 points, and "recieve" for 1 point when a missing letter costs a point in your class. On a name, list "Muhammad" and "Mohammed" both when they earn the same mark. When either spelling counts, list "color" and "colour" on two lines with the same points.
Write the variants you mean to accept. A grammar blank can list "their" for 1 point. Leave "there" and "they're" off the key when those spellings miss that sentence. The key is the list Graider marks against. A wording absent from the key is one you score at review.
A short phrase works the same way. "The mitochondria make ATP" for 2 points, and "mitochondria produce energy" for 2 points, when that second sentence counts in your class. A one-word line, "mitochondria," for 1 point, is a partial you have written on the key. How to keep a list of accepted wordings steady across a stack is in short answer grading with AI.
Order on the line matters when your key says it does. If "6 x 4" and "4 x 6" both earn full credit, list both. If only the first order earns the point, list only that order. A unit you require, such as "12 cm" for 2 points and "12" for 1 point, is two lines. Those points become the one score for the question.
Partial credit on a fill-in-the-blank item comes from the points you write next to each accepted answer. Full credit is the complete response you listed. A middle score is a shorter or less precise response you listed with fewer points.
On "Name the process that turns a liquid into a gas," the key can list "evaporation" for 2 points and "vaporization" for 2 points. "Water turns to vapor" can be 2 points when that sentence counts as a full answer, or 1 point when you want the scientific name for the full mark. Write the points on the line. The question still receives one score and one comment.
A two-blank item is still one question when you score it as one question. List the pair that earns full credit, such as "noun" and "verb" for 2 points, and a single correct word for 1 point when that is your partial. If you score each blank as its own question, number each blank on the test and on the key. Each numbered blank then gets one score and one comment.
Students write fill-in-the-blank answers on the printed test. You photograph the stack, page by page, in even light, with the full sheet in the frame. Students write in pencil or pen. You confirm the roster match, then review before you hand anything back. How photographed handwriting is marked against a key is in how to grade handwritten tests with AI.
Review shows one score and one comment per question. Keep a score when it matches a line on your key. Override it when you disagree, and edit the comment. Toggle feedback on when the comment should appear on the PDF or the email, and off when you will teach that blank with the class.
The only flag is the Low confidence badge on an answer Graider found hard to read. Open every badge. 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.
Fill in the blank grading with AI marks paper tests where students write a word, a name, a date, a formula, or a short phrase on a 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.
Upload the test and the answer key, photograph the stack of student papers, and review the marks before handback. The key lists each accepted answer with its points. Graider applies that key to every paper. You confirm the roster match, override any score you disagree with, and keep the pen.
They come from the lines you write on the answer key. List each full-credit wording, each spelling variant you accept, and each partial-credit wording with the points it earns. Score a wording that is absent from the key yourself at review. When you accept it, override that paper and list the wording on your key for the next test.
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.