Data Input Process

Experts who represent the relevant mentality segment perform data collection. All experts must be aware both of the proposed research subjects and descriptors employed for their characterization.

Data input is preceded by filling out a form that contains personal info: name, age, gender, expert category. This info is further used for the logical filtration of the sample at data processing.

A further data input consists in subject estimation in terms of their properties, as well as a subject's desirability for experts under given conditions. In spite of the fact that subject attributes/properties are the first to be operationally estimated (since a subject's detailed estimate in terms of its characteristics has to precede its desirability evaluation, which is final), it is more efficient, however, to start working with the data input algorithm from the discussion of a subject's desirability (ratings). This has to do with the fact that we use the original data input method conventionally labeled "Ladder-scale", which can be operated in two modes: "linear" (significance) and "non-linear" (characteristics input).

An expert performs evaluation of the subjects on the basis of the scales selected. It consists of 3 stages:

1. An informant is offered all research subjects in a single list that he must split into categories, according to the poles "desirable-undesirable". The process is visually presented as dragging the subject names into one of the two polar lists. Dragging is performed either on the keyboard or a mouse-type manipulator (by drag-n-drop procedure, which is most natural for a person). Preliminary subgrouping sufficiently eases the following process of ranging.

2. If the research subjects are multimedia materials, the software makes them readily available as part of its interface.

3. Research subject ranging is done according to the degree of desirability within each of the two lists.

4. Further, depending on the interface for subject desirability input (specified at project development stage), the expert is offered either "triples" of subjects in the order they received at ranging, starting from a neutral one; or a form for entering numbered desirability values. Subject triples are represented as follows:

And so forth. Here is the -th research subject; - a standard scale length on the monitor screen .

Such a representation allows one to indicate the proximity or distance of the "average" subject to the listed neighboring values of a given characteristic or motive. It is done in a natural way (dragging the subject name with the "mouse"). According to the algorithm proposed, the qualitative relations of subject similarity are transformed into a universal metric (quantitative) scale with the use of recursion relation. In fact, we continuously recalculate the marking value of the following scale on the basis of the previous scale.

Desirability (priority) of the subjects is ranged under conditions specified in the project.

Subject estimation by their attributes/characteristics  (scales) is done similarly to the desirability evaluation ("ladder-scale"), but scale non-linearity for the subjective estimates of characteristics is also taken into account.

After one expert has completed data input, the next expert is automatically proposed to enter his data. As soon as all experts have completed that questionnaire, data from "Tester Modules" is forwarded to the "Basic Module"  via intermediate information carriers (floppy-discs, CDs, etc) or network communications (local nets, Internet etc.), where it is merged with the project database.

EXAMPLE

Main | Contents | Previous | Next