Bazis Intelligence Group monitors all aspects of marketing research process. Quality control of primary marketing information acquired through field work (the inspection of aggregate samples) is conducted in two stages and in accordance with all ESOMAR standards.

Stage I. Quality control during data collection

During the first stage, supervisors revisit 10-15% of the respondents, whose names and addresses have been recorded on the checking list.

• Prior to any fieldwork, interviewers are instructed as to the specific nature of the research.

2.  Piloting

• The length of questionnaires is verified. This is later compared to the average intended interview length.
• Observance of the survey procedure is verified: ways of formulating the interview questions, the use of question cards, interviewer courtesy, etc.

 3.  Screening of received surveys

• The date and time of questioning is verified.

• A supervisor compares the valid social-demographic characteristics of the respondent with what the interviewer recorded.

• The respondent's remarks and comments about the interviewer are studied.

• Some of the questions on the form are put to the respondent more than once. The collected data is compared with the interviewer's results.

• Careful review of the forms unmasks fabrications. The interviewer almost always delivers written answers to open questions. The falsified forms are usually internally contradictory and cannot be regarded as merely a sign of distortion.

Sometimes surveys are conducted in places crowded with people (for example, in the buildings of Moscow's malls), which allows a supervisor to directly control interviewers.

According to instructions, the interviewers are required to inform the supervisor of their intent to go to the place of the survey. This is usually done by phone.

If an interviewer conducts a survey but does not inform a supervisor beforehand, the forms are treated as invalid and excluded. A supervisor also needs to know the starting and ending times of the interviewer's work.

In regions with a high degree of phone coverage, such as Moscow and St. Petersburg, the interviewer's control can be conducted by phone, carried out by a telephone poll.

Random re-questioning of respondents reveals falsifications, and additional indirect questions are used to evaluate the quality of the interview conducted.

Stage II: Quality control during data analysis

Quantitative procedures for data analysis enable researchers to assess quality of data. The deviation of a variable's value from the aggregate average is the first warning sign of inaccurate results. Putting aside insignificant deviations, the average values of one variable by different interviewers need to coincide.

For example, different interviewers would receive strongly varying average income values on a surveyed aggregate of respondents should be the least likely outcome of a reliable survey. Naturally, this doesn’t apply to quota sampling when interviewers are sent to conduct surveys in disparate socioeconomic regions.