Coding and Return Values

Most of the respondents' answers are stored numerically. This section explains the coding of the attribute values. It depends on the question type which values (encoded attributes) a variable (characteristic) may contain.

Tip: A complete overview over your answer codes can always be found in the Variables Overview.

• Empty data field The attribute was not prompted and/or the participant has left the field empty (possible with text entry fields, cloze-texts and additional text entry fields in a selection)
• -1 through -3 A residual answer (e.g. “not applicable”) was chosen in a scale, a selection question, a semantic differential or a text input.
• -9 The attribute was prompted, but the participant did not give an answer. This is not possible with multiple selection (where no answer is also an answer) or text entry fields (empty entry fields are saved as empty data fields – see above). For two-sided sliders, this code has only been available since 10/2011.
• -8 An erroneous value was sent to the server. This may happen with transmission errors or if a participant deliberately manipulates the data that is sent back to the server (e.g. in a scale with the options 1 through 5 the value 6 is transmitted).

Note: Valid answers are always coded with a positive value (>0) in the data set. This encoding can also not be changed (especially to exclude overlaps with invalid values) – if it should really be necessary to store a transformed value in the dataset (e.g. -2 to +2 for a five-level scale), it can additionally be stored in an internal variable using put().

• 1 through n

The minimum of an n-point scale will create the value 1, the maximum the value n. Steps in between will correspondingly create the values in between.

The output value is independent of the labeled numbering. If the minimum is labeled e.g. with the school grade “6” (the maximum being “1”), the stored value will still be 1 (correspondingly 6 for the maximum).

For sliders, the value is stored between 1 (minimum) and the value specified as differentiation in each case. By default, this is 101.

• 1 through n

The most negative pole / minimum will create the value 1, the most positive pole / the maximum of an n-point semantic differential will create the value n. Steps in between will correspondingly create the values in between.

The stored value is independent of the labeled numbering (e.g. if the questionnee is shown a scale from “-2” through “+2” – it will be stored as 1 through 5). The negative pole will always create the value 1.

• 1 through m

The value stored will be the Item ID (visible in the Variables Overview or as the value in the square brackets at the list of questions). In case the order of the options shown to the participant should differ from the output values, you may vary the item position in the “edit question”-area or state the item ID order in the control code, e.g. question("AB01","3,2,1,4,5,-1");

As with a dichotomous scale (“not selected”/“selected”) the values for a multiple choice are saved as 1/2 for each item.

• 1 Item not chosen (“not selected”)
• 2 Item chosen (“selected”)

Dichotomous data will be recoded to false/true if using the GNU R import:

• FALSE Item not chosen (“not selected”)
• TRUE Item chosen (“selected”)

In case you have given the option to enter an additional option through a free text entry field, this value is saved in a separate variable (see the Variables Overview).

Bear in mind, the value -9 (“no answer”) is not possible with multiple choice questions.

The entered text is taken over directly; however, some characters (e.g. for column and case separators in the CSV file, see Download Data: CSV) may be replaced by exclamation marks (!) or pseudo tags (e.g. <comma>).

The codes -1, -2 and -9 do not exist for text input.

Of course, numbers (e.g. age or year of birth) can also be asked as text.