Open DilekIkiz opened 4 years ago
I think it would need a bigger overhaul than what you proposed and came up with the following design types:
single observation/cross-sectional
single observation/cross-sectional but with retrospective/prospective variables
repeated cross-section/trend design (e.g. what Allbus does)
longitudinal designs they have the following subtypes:
repeated observation of the same units
repeated observation of the same units with panel variables (most often data is not collected in a pure panel design)
What about experimental manipulation? The domain object is called SurveyDesigns after all.
At the moment I would shift the focus from the wording towards specifying a more complete list of design types and in a second step agree on the wording. Furthermore, as both, the object is called SurveyDesigns, and a survey can have multiple of these characteristics, we should probably allow for a list instead of a single choice.
This is a good suggestion!
@DilekIkiz we have to use da|ra terms, these are:
Unfortunately, they are not really adapted to qualitative data ....
But we do could allow for a list of designs ...
@DilekIkiz please give feedback
Check list with ddi.
Sorry for my late reply. The issue has completely disappeared.
My irritation is not about the longitudinal type, but about the cross-sectional type. In qualitative studies, sampling is not (always) designed to represent a cross-section in the sample. Therefore, the term cross-section in itself leads to irritations for me and also for data providers.
@rbirkelbach had research done on the different study designs. According to this : "The cross-sectional survey refers to a point in time or a short period of time in which a one-time survey of the properties of N examination units is carried out". So the unit of time is central, as I understand it. Also in da-ra (and DDI) the differentiation is based on the time unit: "timeDimensionType"
So it is not really the spectrum of "(survey) designs" that is listed here.
Overall, according to this logic, the qualitative data would belong to the cross-sectional designs.
@DilekIkiz is currently discussing the topic "metadata for qulitative data" as part of the Konsort measure Qualidata Network, so we are postponing this issue until we have concrete results here. The next meeting concerning metadata is on the 30.06.
In the field „survey designs“ the two options Panel or Cross section do not adequately reflect qualitative research designs. They are not understood or irritated.
Panel are still conceivable, however, the term "inquiry at several points in time" would probably be more appropriate. On the other hand, cross sectional design completely gives rise to an irritation for the data provider. Qualitative studies are mainly designed as "Einmalerhebungen".
In consultation with @danbfdz, I removed this field in the overviews of dps2018. In my opinion, there are three possible solutions for MDM: