Edit the questions of a quiz

Catalogue games are quizzes in disguise: whatever the staging, the player moves forward by answering questions, and the score sent to your reports and to your LMS is the score of those questions (the share of questions right on the first try; speed, streaks and lives earn game points, kept apart). For these games, the editor has three spaces instead of the usual modes:

• Questions: the questions themselves, their order and draw, and the shared answers;
• Appearance: images and interface texts, with the live preview;
• Rules: timer, jokers, passing score and number of questions per game.

The Brand space (colours, font, logo) comes with them, as for every game (see Edit your draft). For these games, the Advanced mode is kept for the Maketools team, and there are no Content and Screens modes. Games that are not in the quiz format keep the editor described in Edit your draft. Once the fork is deployed, the game's report shows the success of each question (see read a game's statistics) and, in an LMS, each question comes through as an interaction (see LMS integration).

These spaces are those of games with the Editable badge. A game without this badge only has the Brand space (see Games that are not editable).

Fields that stay within their limits

When the game declares its fields (AI Summit does), every field of the Appearance, Rules and Questions spaces has a control made for what it expects. Nothing is typed outside what the game accepts, and no identifier is ever entered. The table lists every control; AI Summit uses texts, numbers and choices:

Fixed lists: when a game ships a list of items (characters, events…), you edit their texts and images, without adding or removing any; the fields that point to them offer them in a drop-down list. No published game ships one yet.

A value outside its limits is flagged under its field as you type, and the server runs the same checks on every save (see Checks on every save). Games that do not declare their fields yet keep spaces drawn from their configuration, checked on save.

The Questions space

The list shows every question in the order of the content: its number, its kind, its prompt, its skill, and what to review. Switch to Table to change the difficulty, the weight or the opener of many questions at once. Search looks into the prompt, the answers and the explanation; the filters keep one kind of question, or the questions to review.

To reorder, drag a question, or press Alt+↑ / Alt+↓ on it. Clear the search and filters first: the order is the one of the whole list.

New question creates a question from a template of the kind you choose. Only the kinds the game knows how to present are offered:

No catalogue game offers the short answer yet.

The sheet of a question

• The question: the prompt, an optional situation and an image.
• The answers, according to the kind: the right answer is ticked and marked "Right answer". Answers can be grouped (named groups, such as the zones of a slide).
• After the answer: the explanation every player sees, and an optional hint.
• Score and draw: the skill, the difficulty, the weight (1 to 5: a question of weight 2 counts twice in the score) and whether the question can open the game.
• Staging: what this game needs to present the question (never used in the score).

Each change is saved automatically and can be undone (Ctrl+Z). The editor never lets you save a question without a right answer: a "several answers" question keeps at least two right answers and one wrong answer, an order lists every item once.

Points to review

Some checks never block saving but flag what makes a question ambiguous or easy to guess: an empty prompt or explanation, the same answer twice, a right answer much longer than the others (more than 15 % longer than the longest wrong one), "all / none of these", a negative prompt, the same prompt twice, or a right answer always at the same place. They are listed on the overview of the Questions space and on the sheet of each question.

Order and draw

• Order: the order of the list, shuffled for each game, or by increasing difficulty (easy, then medium, then hard). In a shuffled order, one of the questions marked "can open the game" comes first.
• Questions per game: all of them, or a draw of a given number, optionally balanced by skill. The server replays the same draw to compute the score.

Shared answers

For games that classify (for instance "AI" / "Not AI"), single-answer questions can share the same answers, written once. On a question, choose Share these answers with other questions; the other questions can then use the shared answers. A shared answer that is the right answer of a question cannot be removed.

Import and export as CSV

The Import / export (CSV) menu downloads the questions, or a blank template with one question of each kind, and imports a file edited in a spreadsheet. The file has one row per question and these columns (header line in English):

How to write the answers:

• One answer and several answers: one answer per line; a line ## Name opens a group. correct is the letter of the right answer (B), or the letters separated by ; for several answers (A;C). A one-answer question without answers uses the shared answers: correct is then the letter of a shared answer. Letters count every answer line, an empty line included (an answer still to write): the right answer never moves to its neighbour.
• Order: the items in the expected order; correct stays empty.
• Pairs: one pair per line, left => right; a decoy on the right is written => right.
• Number: correct is 42, or a range 40..45.
• Short answer: the accepted answers, one per line, the model answer first; correct stays empty.

Example:

id,kind,prompt,answers,correct,explanation,skill
q-1,single,Which task suits a language model?,"Computing the payroll
Drafting a first email
Signing a contract",B,Drafting is a good fit.,ai.usage.good-fit
q-2,numeric,What share of the panel answered yes?,,40..45,42 out of 100.,data.reading.proportions

The import replaces the list of questions, in the order of the file. Nothing is written until you confirm: a refused file lists every problem with its spreadsheet row and column. A question that keeps its id keeps what the file does not carry (image, staging, feedback of its answers), and an answer whose text is unchanged keeps its id, so the reports per question stay comparable. The order and the shared answers do not change; the draw too, except that "all the questions" becomes a draw of 100 when the file has more (a game shows 100 questions at most). Commas and semicolons both work as separators; the file is UTF-8.

A text that starts with =, +, - or @ is exported with an apostrophe in front ('=…), so that a spreadsheet shows it instead of running it as a formula. The import removes that apostrophe. Numbers such as -5 are not changed.

Before publishing: the game played through

When you publish, Maketools first plays the game with your draft: a perfect play and random plays, over several draws. A question that cannot be played, or a maximum score that cannot be reached, blocks the publication; the editor lists each problem with its field and a Fix button, and your draft stays saved. See the version history.

Translate the questions

Questions are translated like every text of the fork, with the translation file of the Languages tab: see Translate a fork.