Response domains are not limited to one per question, as some questions may offer the choice of how to answer (for example having both a code list and a numeric answer response; or a code list with a Generic Response)
Example 1 Questionnaire: ALSPAC All Around Me
Archivist view (ALSPAC_03_aam):
Example 2 Questionnaire: ALSPAC My Teeth
Archivist view (alspac_99_mt):
The question literal itself must contain any important context for answering the question, particularly relating to the who, when and where.
In Example 3 below, the important context is the study child (who) and evenings (when).
Example 3 Questionnaire: NCDS Parental interview form
Archivist view (ncds_74_pq):
An exception to this rule, is when the text that provides the context is entirely contained within another question item, which the question is dependent upon. This often happen when the text is in a condition based on the first question. If you look at Example 4, the question asks “Did you knock any teeth out?” and then the next question asks “How many did you knock out?” which is inside a condition of “If yes”. The second question lacks context, but as it is within a question which contains the context, this is acceptable. Adding text to the second question would violate the Prinicple of not improving the questionnaire.
Example 4 Questionnaire: ALSPAC Teeth and Things
Using question items
Question constructs house question items, and therefore you do not need to enter identical questions twice, it's very rare where the question items are used more than once, as the question components are specifically built for the question items made. For example if the identical question is used in the questionnaire twice and produces the same data then the question item can be reused. However, they will need different question constructs with unique ID labels, as these represent the positioning in the questionnaire.
Example 5 Questionnaire: SWS Initial(SWS)
If presented with a sub-question situation (see Example 6) then you must concatenate the question literal to produce the most concise method of documenting the question and provide the essential contextual information. The question literal should be made up of the common text for all questions followed by the sub-question text.
You can see from Example 6, the question literal here is, "How would you assess the health of your child nowadays?" and concatenate texts "in the past month" and "in the past year", to produce the two seperate question items. The response domain will be the four categories you can select in how to answer "1 - very healthy; 2- healthy, but a few minor problems; 3 sometimes quite ill; 4 almost always unwell". The catergories in the response domain have a numerical value linked to each, indicating a code list, and therefore the literal text be recognized as a question item. This flags up an important principle, not be seduced by the presentation of the question, but how you interpret the information presented following our entry guidlines, as Example 6 could wrongfully be misintereped as a grid (lacks 25 word count, etc).
In Archivist we have replicated question A1, with two question items, labelled "qi_A1_i", and "qi_A1_ii".
The alternative to a question item is a Question Grids, which is used to enter more complex question situations. See Questions for information about how to decide between a Question Item and a Question Grid.
Example 6 Questionnaire: ALSPAC My Son / Daughter's Health & Behaviour(ALSPAC)
Archivist view (alspac_94_msdhab):
You can see from Example 6, that in Archivist we have replicated question A1, with two question items, labelled "qi_A1_i", and "qi_A1_ii". The key difference between "qi_A1_i", and "qi_A1_ii", is the duration of the question, "in the past month" and "in the past year."
There are some questions which require more than one code list. This is identified by the change in code values as shown in Example 7, below.
Example 7 Questionnaire: NCDS Medical questionnaire 1974
Archivist view (ncds_74_mq):