Questioning What types of questions are there

Grafstat software

After the questions have been formulated, they must be typed. Each question can be assigned to one of the four elementary question types:
  • Choose;
  • Scale (rank scale);
  • Measure;
  • free question.
The assignment to a question type is not arbitrary, however, because the question type is closely related to the answer options (characteristics, characteristics) and also plays a role in the options for subsequent evaluation! Each type of question must meet certain criteria, which become more and more stringent as they are selected by measure.


When making a selection, the various answer options are named, which create clear assignments, i.e. must be selective. The selection is available in two variants: multiple choice and single choice.

Multiple choice

Multiple choice questions are a special form of multiple choice questions in which multiple answer options can be ticked. Multiple choice is the most undemanding type of answer. However, it causes a lot of effort and headache to find the possible answers or to limit oneself to certain aspects. With the evaluations, the possibilities hardly go beyond simple quantitative representations. For larger projects, a preliminary survey with free questions can significantly reduce the uncertainties.


Own vehicles:

A) bicycle
B) moped
C) motorcycle
D) car
E) other ___________

GrafStat automatically precedes the answers with the answer abbreviation, in this case capital letters. Up to 26 answer options are allowed.

Single choice

Single-choice questions are a sub-form of multi-choice questions in which only one answer can be ticked. In the case of a single choice, the answers must not overlap, they must form unambiguous classes. In addition, the entire answer space must be covered, i.e. it must be possible to tick one answer in each case. In many cases, an answer such as "Other" or "Other" should therefore be in the answer list.



a) female
b) male

GrafStat automatically precedes the answers with the answer abbreviations, in this case lowercase letters. Up to 26 answers are allowed.


In the case of questions of scale, the answers have an internal order, a predetermined order. The order is also referred to as a ranking (ordinal scale). This type of question is recommended, for example, to record moods and expressions of opinion. For a better understanding, the individual values ​​are provided with terms. Often only the two outer extremes are named. GrafStat then displays the scale in the questionnaire in a single line, with the extreme terms to the left and right of the scale.

Scales can have an even number or an odd number of values. Scales in particular have no center and therefore no neutral value. You can also create symmetrical scales, for example from –2 to +2. Such scales are suitable if the question allows a negative / positive assignment. In empirical social research, great efforts are made to construct meaningful scales for characteristics (e.g. intensity of interest in politics).


How much do you like chocolate?
very much 1 2 3 4 5 very reluctantly

What grade do you have in mathematics?
very good 1 2 3 4 5 6 insufficient

Symmetrical scale? If negative and positive should be clearly recognizable in a rating, then you can use a scale with a negative starting value. When making this decision, however, you should take into account that the shape of the scale has a certain influence on the response behavior. If the scale is structured symmetrically, with a "neutral middle" (e.g. -2 -1 0 1 2), undecided (neutral) answers are possible. If you do not specify a middle, i.e. use a scale with an even number (e.g. 1 = more likely to agree to 6 = more to disagree), then ask the respondent to make a clear decision.

Name all values ​​or just the end values?

The appropriate formulation of all intermediate values ​​sometimes causes difficulties. Think about whether you have really found clear and orderly formulations. It is often easier and better to label only the two extreme values. In the case of symmetrical scales with a negative initial value, naming the extreme values ​​is often sufficient. Also consider the appearance on the questionnaire. With the reduced lettering, the cross boxes appear in a row, with the names on the two flanks.

Couple scales?

In questionnaires there are often whole blocks of questions with the same scale. For such cases you can link the questions with each other in GrafStat. In all forms, the questions appear one below the other on the left and the checkboxes to the right of them. The labeling of the scale values ​​is simply inserted above the cross box.

Choose a scale, though
  • ... a feature can be classified according to size.
  • ... it's about assessments.
  • ... something should be rated.
A scale is unsuitable
  • ... when determining whether something is present or given (yes / no).
  • ... for assignments (e.g. school attended: Hauptschule, Realschule, Gymnasium, Gesamtschule).


Measurements make the highest demands. Measurement questions require a unit of measure. Dimension numbers must have a zero point and an equal distance between the possible measured values. Many physical units of measurement meet this requirement (e.g. length in cm, weight in kg or time in hours) (interval scale). GrafStat requires the specification of the unit of measurement. For the purpose of input security, the lower and upper limit and the number of decimal places are also recorded.


Weight: ___ kilograms

Age: ___ years

TV consumption: ___ hours / day

Minimum and maximum value

The two values ​​are used as a control. As soon as the questionnaire is entered, typing errors or intentional incorrect entries are rejected by specifying the minimum and maximum values. Set both values ​​to the expected limits. Note that higher or lower values ​​can no longer be recorded.

Note: If you want to work with comparative figures from other studies or statistical yearbooks in which groups were formed (e.g. age: 16-19, 20-29, 30-39 years, etc. or income: less than 500 euros, up to 1000 euros, etc.), so you should also adapt your survey to these groups. This facilitates the comparison during the evaluation and is even necessary e.g. for data handling functions such as weighting. In this case you should type the question as a single choice.

Free question

Free questions (open questions) have a special position. They cannot be statistically evaluated automatically, i.e. they can be typed in, but must be categorized and counted "by hand"! However, you cannot completely do without free (open) questions, because they perform helpful tasks. In a preliminary survey, they are very well suited for the construction of a questionnaire in order to capture all the interests of the target groups. Information can also be obtained via the free (open) question, which opens up completely unknown or previously neglected aspects.

What is a free question?

A free or open question is a question for which the answer can be freely formulated. No pre-formulated answers and no checkboxes appear on the questionnaire, only a number of writing lines that you have specified. A multi-line text field for text entry appears in the HTML form of the survey for the Internet.

Note: Free questions cannot be counted and evaluated statistically with GrafStat.

When do you use a free question?

When you don't know the possible answers but want to find out, when you don't want to limit the respondent, or when you need lively comments or the like. The free question can be helpful if you are conceiving a new questionnaire and in a small preliminary questioning, for example, to find out the answer options for a single or multiple choice or if you want to get other suggestions.

Evaluation of free questions

A free question cannot be automatically counted by the GrafStat software, but has to be evaluated by hand. Since the free question does not provide any quantitative data, the evaluation will consist of collecting, arranging and categorizing the answers and, if necessary, counting them.

With GrafStat, the texts of the free questions can be recorded during the data collection. During the evaluation, all answers to a free question are presented together in a text document and can be transferred to a word processor or another program for further processing.

If the free question is part of a preliminary survey, then you will
  • group or categorize the answers,
  • to summarize elective answers,
  • formulate the items.
In the questionnaire for the main survey, instead of the free question, insert a multi-choice question (single or multiple choice) with the most frequently given answers to the free questions.

Formulate answers

When formulating the answers, you should carefully follow the guidelines that also apply to the formulation of the questions. Here, too, one should orientate oneself on good examples if possible.
  • short answers as possible;
  • look for clear and accurate terms;
  • eliminate duplicate and similar answers;
  • Weed out unnecessary answers.
A low number is supported by the fact that the respondent will find it more difficult to cope with a large number of answers than with a smaller number. In addition, especially in the case of smaller surveys, the frequency of individual responses quickly tends to zero if there are too many possible answers. Statistical evaluations then become quite problematic.