The tangle of questionnaire design options

People who do research must have a book about how to design questionnaires. Looked at, found that "questionnaire design process", "Problem type and Application", "a qualified questionnaire assessment criteria" and so on (the book by the Mechanic). There is no "design of the choice", let a person tangle.

Textbook does not seem to, then look at the field of practice it.

Barbara Minto (Barbara Minto),  first female consultant, presented a very important principle in the pyramid principle (the Minto pyramid principle): Mutual independence, complete exhaustion (mece,mutually Exclusive collectively exhaustive). It means that effectively grasp the problem and solve the problem, first of all need to classify the problem, this classification must do not overlap, do not omit.

Although this is a principle in the area of management consulting, it is actually common in the design of options. The question we encounter is, what frameworks can help us to make sure that there is no overlap or omission?

It is the main intention of this paper to try and refine this frame of thinking.

Refer to "option", must be related to the choice of questions, the choice of the classification of questions are: Radio/multiple selection, closed/semi-closed. The former is related only to the number of options, which involves the design of the option, so this is mainly discussed in the latter category.

First look at the closed, rough next definition.

Closed: Options on these, you can only choose among these options, such as: Have you eaten? Eating, eating, not eating.

Imagine, the option is only "eating", "finished", what will happen? Yes, people who don't eat have no choice.

So, with an option design rule "option must cover all possible aspects as much as possible." In other words, when designing options, "You must know all the possible eating states."

"Eating status" includes "eating" or "not eating". In the "Eat" inside add a Time dimension, can be divided into "eating" or "finished." As a result, possible eating conditions include "eating", "Eating Out", "Not Eating" (and, of course, a number of dimensions). The option design of closed-end problems is simpler, as long as the logical analysis can cover all possible situations.

Some examples:

Look at the semi-enclosed, also a rough definition.

Semi-closed: In addition to the options given, you can also choose other, such as: What do you think the world is? Black, white, other (please specify). You can choose the other and then indicate the ash.

Why is it half closed?

The festival is coming, you want to send a gift to someone, after shopping malls found that pocket money, you can buy a, B, C, D four things in one, you think the other party may like a, C, D, but do not know what to prefer, do not know whether they do not like it (perhaps like other), so on the card, wrote a semi-closed problem, Let the other person fill out. Ask "What is your favorite gift?" , give the option "A, C, D, other (please specify)".

To do these, nothing more than two roads, one is to find something I can buy, in the selection of the other side may be like things, the other is to delineate the other may like things, in the selection I can buy.

Whichever way you go, know what the other person might like. However, the problem is to ask each other what they like, haven't asked, how do you know what they like?

So bite the bullet, pick a few things, add another (please specify), ask each other. The other side may say: "Other, annotated E." This e is out of the plan, and the result may be that you can't find a running leg or a budget.

What's worse, the other party may only choose the other, did not specify the content, this time is equal to white ask, only know that you choose each other do not like, but still do not know what to buy something as a gift.

Ideally, although there are "other", but the other side in the ACD made the right choice.

To each other in the ACD to make the right choice, ACD must include the other's favorite choice. In other words, the ACD itself needs to contain what the other person likes.

As mentioned above, when designing options, we are not sure what the other person likes, so we ask ACD to cover as many things as they like.

This is common with the option design rules for closed issues.

The option design of closed-end problems, mentioned above, "As long as the logical analysis, can cover all possible circumstances, can be", then the semi-closed problem can also be logically inferred? The answer is yes.

Just because the conventional logical framework (degree/order/scope) does not apply, we need to find the appropriate analytical framework for logical reasoning in the framework.

Share a few very useful framework ideas:

The above mentioned ideas, in addition to being used in option design, can also be used in model construction.

"Touch comb."

"Product Deconstruction"

According to the Convention, multiple-choice options, should be through open questionnaires, based on user feedback, but the user feedback and option design, which has a very important link, that is: based on some kind of framework ideas, combined with the product/service/theme of the structure, summed up user feedback, Can finally design a set of effective options. in this context, the design of options and model indicators are in fact interlinked. This article sums up sharing is exactly what this link should do.

Market research is an interdisciplinary interdisciplinary science. The model or framework itself is not the most important, implicit in the thinking behind and the expansion of the application of ideas is the most worthy of our time and energy to learn the place. In the process of writing, once again to this sentiment deep.