There is significant variation between the structure and the content of the protocols. Some regions provided more detailed information about the candidate’s background, others - more general. In some cases in one protocols there was information about more than one candidates. That is why we have chosen the method of manual coding of the protocols. For coding we used the system of biographical data collection BORIS. The coders were instructed to extract relevant text parts from the resolutions of the qualifying board of judges and fill subsequent fields.
Despite the regional variation of the approaches to write protocols, in most …show more content…
Basic data: year of appointment, region, level of the court (court of the peace, district court, regional court, and arbitrazh court), position for which candidate apply (judge, vice-chairman of the court, chairman of the court).
2). Biographical data: birth year, birth place, university graduated, year of graduation, mobility (the coders were asked to evaluate from biographical data whether the candidate moved to another place (different from birth place) for study or work) and whether the candidate’s current locality differs from the locality of perspective place of work.
3). Occupation data: place, position and period of work at each place.
4). Outcome of candidate consideration by qualification board: whether the person was recommended, why (if available), the voting results and credibility checks.
All initial data was inputed in natural language for this reason we used OpenRefine as a primary tool of data cleaning and preprocessing.
Dependent and independent variables
The dependent and independent variables are displayed in the Table 3. Recommendation of the qualification board is dependent variable that coded “1” if the candidate received recommendation and “0” if the application was