Joseph Rézeau

Vous devez en effet avoir une ancienne version de cette activité H5P. Comme on le voit sur ma capture d'écran, le score est maintenant affiché directement à la suite de la barre de résultats (verte) : ici 1/2.

1.- S'agit-il bien de l'activité "Quiz (Question Set)" ou d'une autre activité H5P, comme par exemple Course Presentation ou autre chose ?

2.- Est-ce bien la dernière version à jour de l'activité Question Set (1.16.2) ?

3.- J'ai fait un essai (avec Quiz Question Set version 1.16.2) et ça marche sans problème. Il faut bien faire les 2 manipulations suivantes:

a) pour chaque question, dans la section Paramètres comportementaux, décocher la case Activer le bouton "Afficher la Solution"

b) pour le quiz lui-même, dans la section Quiz terminé, décocher la case Afficher le bouton "Solution".

Sur ma capture d'écran le message "vous avez moins de 75%" provient de la section Opacité des étiquettes (traduction fantaisiste de Overall feedback !) Définir un retour personnalisé pour chaque tranche de score, où j'ai défini 2 tranches de score. Je suppose que vous avez fait de même pour votre message "Vous n'avez pas réussi ce questionnaire." En revanche je ne vois pas comment ni où vous avez défini vos messages "Vous avez obtenu 5 point(s) sur 31 ni "Refaites un essai!".

@Luigi,

I had based my reply on your original post, in which you compared a quiz made of 100 short answer questions with a quiz made of 100 multiple choice with multiple answers questions. But then, in your second post, you changed the terms of the comparison to multiple choice single answer questions compared with multiple choice multiple answers questions which is a different matter.

Sorry I had not realized that "switch" which makes my reply irrelevant to your problem.

However I support the replies already given by the other two forum members who replied to you.

Hi Luigi,

I'm afraid your experiment is flawed from the beginning. You are comparing the results of 2 totally different types of questions: an open-ended type (short answer) and a closed one (MCQ).

Unfortunately you do not give any concrete example of the questions you asked of your students, but it does not matter too much. Let's take those 2 question types one after the other.

Short-answer

This type of question requires the respondent to enter the exact wording of one (or more, if you provide correct variants) answers. If you asked such questions from respondents presumed ignorant of the domain, it is only to be expected that "they got 0 points on each question, and 0 points on the whole quiz".

Multiple Choice

I won't deal here with the matter of MCQ with more than one correct answer nor the matter of positive/negative marks. i can only say that when asked to select from a finite set of answers, the laws of probability make it very unlikely (and indeed impossible) that the mean score of the respondents will be zero, like the mean score of the short answer quiz.

Your questions:

"why students got 0 points (on average) on the first quiz?"

I have answered that question above; that score was only to be expected.

"and they got 20 points (on average) on the second one?"

As I explained above, there was almost no chance that the mean score on a quiz with that type of questions would be zero. You were wrong to expect the same kind of mean score when using those 2 totally different question types (open-ended/finite vs closed/finite). I am not an expert in maths or statistics (unlike my distinguished colleagues Jean-Michel and Dominique). There is probably some maths formula into which you could enter your scoring data (2 correct answers marked 50% each + 4 incorrect answers marked -25% each) which would give you some expected mean score (for a quiz with a large number of such questions answered by a large number of respondents).

I have just conducted a minimal experiment with just one MCQ question using your scoring scale. I have set the question to shuffle the answers and - in question preview - I have systematically checked the first 2 answers, which were different for each try. My first set of 10 tries yielded a mean score of 7.5% and the second set of tries a mean score of 27.5%. This little experiment has no statistical value, it just goes to show that is quite possible to achieve those mean scores purely by chance.

Merci, Mary, de nous avoir signalé ce plugin que je ne connaissais pas. C'est un peu compliqué à mettre en œuvre, mais on y arrive. En fait, ça permet d'insérer des questions de la banque de questions d'un site moodle (restreinte au cours courant) non seulement dans des pages de type "contenu" d'une leçon, mais dans n'importe quelle page d'un cours dans laquelle on a accès à l'éditeur HTML (Atto). Et ça fonctionne très bien...

Malheureusement, en ce qui concerne l'activité Leçon, les scores des réponses données aux questions ne sont pas décomptés dans le score total de la leçon. En somme, l'étudiant répond "pour du beurre" à ces questions insérées.

Sur la page de ce plugin, Tim Hunt écrit : "[for example] a student asks in a forum about some tricky grammar concept, and you have a question that illustrates your reply, well now you can include that qustion in your reply, so the student can immediately test their understanding."

Traduction Google: "[par exemple] un étudiant demande à un forum sur un concept de grammaire délicat, et vous avez une question qui illustre votre réponse, eh bien, vous pouvez maintenant inclure cette question dans votre réponse, afin que l'étudiant puisse immédiatement tester sa compréhension."

C'est en effet un très bon exemple de suggestion d'utilisation de ce plugin.