Frequently Asked Questions

We suppose here that we have a shop selling online women's clothing.

We want to organize a survey in the form of test to better understand the behavior of our clients in mater of dresses. We would like to highlight 1, 2 or 3 types of dresses that would win their vote. We could refine our offer and also provide "basic" models throughout the year.
We take the opportunity to better understand their tastes in colors, but also in printed material, tissue materials, and length.
We shall can refine these data in matter of season, opportunities to wear dresses, etc ...

These first points already define the type of questions we will ask. We plan to ask 10.
They will therefore relate to dresses styles, to the looks, to materials, printed, opportunities to wear dresses, the colors ...

Since this is a test on fashion and on the dress behavior, we chose to determine for each player a profile / type "fashionista".

"Tell us about the dress that would make you crack and discover what fashionista family you belong".

It will appear in the final part of the test, when it has answered all the questions.

Here are profiles or groups / types which I create:

- The family of passionaria, who will be my majority group A (all the answers = A or at least 6 answers A)
- The family of romantics, who will be my majority group B (all the answers = B or at least 6 answers B)
- The family of genuine rebels, my majority group C ( all the answers = C or at least 6 answers C)

- The family of sportives comfortable with themselves, egalitarian group A / B (same number of A and B and greater than the number of C)
- The family of mysterious elegant, egalitarian group A / C (same number of A and C and greater than the number of B)
- The family of Nature women, egalitarian group B / C (same number of B and C and greater than the number of A)

This is where lies the tricky part in designing a test. These profiles / types are called "groups" in the MultiQuiz module and treated by letters: A, B, C ... assigned to each response. It is thanks to this classification that the MultiQuiz module will sort and distribute the answers to determine the group to which each player belongs.
When we write our answer choices for each question, one must always keep in mind our determinants groups (here: A, B, C).

- Consider an example with the first question of the quiz:

Question: In general, you wear dresses:

Choice of answers:

-Very often (A)
-Rarely, I prefer pants (C)
-Regularly (B)

For this test with three majority groups, so I have to plan for each question at least 3 different responses, one for each group. This is the mandatory minimum for my test is consistent and relevant.
It is when I fille out the form for each response in MultiQuiz that I will indicate to which group it belongs.

Egalitarian groups are not mandatory, it is an option we choose in the Configuring of treatments.
These are the results of the players straddling two groups: they have eg 5 answers A +  5 answers B or 5A + 5C, 5B and 5C + or 4A + 4B + 2C or...

There remains a fourth egalitarian group where I could find some players, those that could equal between answers A, B and C.
This can happen in this particular test because I proposed to several issues 1 choice of 1 to 3 answers. So I can have a lot more than 10 responses to 10 questions. One can avoid writing a profile for this group by deciding to treat it such as group A/B, B/C or A/C, or by doing a test with an even number of questions and by permitting only one response by question. Then it is mathematically impossible to have a equality between A, B and C.

For this test, I decided to treat equality of Answers A / B / C as egalitarian group A / B.

I could also decide to bypass the other 3 egalitarian groups. It would be enough to indicate that A/B must be treated as answers A, B/C as answers  B, and so on. At that time, I would only create my 3 profiles / majority type. This option would shorten the test preparation time, but it is not always very relevant depending on the chosen theme. In the case studied here, three groups were insufficient to describe the diversity of profiles in fashion.

The test requires work of elaboration a little longer than just a quiz, but it is a valuable marketing tool and very thorough. The MultiQuiz module does all the work of calculating for you. You simply have to choose your options for configure the questionnaires, and devise questions / answers + profiles / type.