The research was conducted in the period July 3rd 2001 - July 23rd 2001, during which thirteen specially trained nonprofessional interviewers investigated the normative documents, and the main problems, concerning Roma education in all regions of Bulgaria. The period was deliberately chosen to coincide with the time when the Bulgarian educational system draws the public attention with its annually held exams for high school and university admissions. Furthermore, the time-period of the research coincided with the expected changes in the government. According to the polls, in some regions the prospect of new individuals entering domestic politics was of greater interest than the problems of this research.

     In the light of the political changes at the time, raising the "Roma question" appeared somewhat provocative, but its rationalization was triggered by middle and low governmental units. From an institutional perspective such units are the regional body of inspectors, the municipalities and the school principals - the notorious educational triangle, where responsibility and valuable information always go opposite ways. Our aim was to understand the extent to which these units were sensitive towards Roma problem, their readiness to change, and willingness to work.

     Pursuing the above reactions, we interviewed 18 heads of educational inspectorates, and/or 28 deputy inspectors, 58 public servants in Roma-inhabited municipalities (heads of divisions, specialists on educational, ethnic and demographic issues), as well as 72 school principals or deputy principals, teachers, and school workers. From all of the above-mentioned personnel no one is of Roma origin; in most regions there are no Roma teachers either.⁵

     Sixty-four schools with a predominant number of Roma pupils were visited. Also, interviews were made with 26 Roma students from different universities, as well as with 12 managers and workers from NGOs involved in the development of educational programs for the ethnic minorities.

     In all the cases the interview was held in the form of a conversation, which gave a certain freedom to the interviewed person in expressing his own ideas, not minding other opinions, and helped them overcome any "politically correct attitudes". In the course of the conversation the language also finds its natural ways and the word "Gypsy" replaced "Roma" - a still unpopular form of expression to identify the minority. Many of the interviewed pointed out that no one, except those directly involved, takes interest in the issue. There are cases of purely hygienic segregation and racist attitudes, so as to "protect" the other children. All of the interviewed officials possessed self-esteem and self-image of a high degree, which not always coincided with the public evaluation of their office. The interviewed would speak of Roma from a managerial point of view. Such approach standardizes the different institutional interests, but also presents contradictory information. However, the aim of the research was not to find the highest level of authenticity, but to survey the reaction to the given problem.⁶

     The same approach should be used in the analysis of the database, the results of which probably would have changed significantly, if such a research were conducted among the Roma community. Moreover, it can be predicted that due to the fact that the interviewers are outsiders, the number of respondents who pronounce their minority origin would have diminished. This is also confirmed by the reaction of most of the interviewed persons, who know how Roma people identify themselves, but they always make the remark "but they are actually Roma." The same is true for the reaction of the educated Roma: "our kin, they think they are... they want to be Bulgarians, Turks, seldom -Wallachs." In the last census some ironic persons once again defined themselves as Latinos, Eskimos, Indians, etc.

     Certainly, the presented information is tentative. An adequate qualitative research is almost impossible and dangerous, given the hidden prerequisite that the self-identification is to a certain extent modeled from outside factors. This effect can be noticed in some of the data concerning schools and regions (check the regional results - the differences in percentages are due mainly to the closeness to the real self-identification).
Most often Roma prefer to be defined, and define themselves, as ethnic and religious majorities in the regions they inhabit, although in these regions they are often unwanted. This holds true for regions inhabited with Christian Bulgarians, and regions inhabited with Muslim Turks. However, in regions inhabited with Muslim Bulgarians Roma traditionally are absent. The few Roma there are rejected by the majority on both ethnic and religious reasons.

     Thus, the research's results, concerning Roma schools, can be analyzed only as relatively trustworthy statistical data; in this sense the research does not guarantee statistical accuracy. Rather the research has a phenomenological value as a document trying to capture, and provide for, dominating arrangements in the definition of a certain reality, Roma schools.