Since this leads to serious errors, many research companies are testing up to 50% of the results interview. This process involves repeated, usually a phone, a survey of a number of participants study. Non-response error is a significant problem, because people who refuse to participate in research is likely to be significantly different from those who went to cooperate. There are several reasons for the refusal to participate in the study. Mainly – the fear of consequences. The study also may be perceived as an invasion of privacy.
In addition, people can be unfriendly configured in relation to the researcher. Error unable to answer. Respondents may be unable to respond if they did not know or forgot, they also may be unable to properly express it. Problems in each of these cases may be even worse if the respondents will invent answers, because they do not want to admit his incompetence, or want to favor the interviewer. It should be remembered that in most cases, respondents are willing to help the researcher and try to give as much information. Contribute to research an error can make the distortion of time, averaging, or omission of information. During the interview errors may occur reluctance to give a precise answer and distortion response. Among the reasons for this phenomenon may be a sense of invasion of privacy, lack of time, fatigue of the respondent, for reasons of prestige and social acceptability of an answer, courtesy, lack of information and style of response.
Input errors, data analysis and interpretation After all the completed questionnaires were delivered to the office of Defense Research Agency, specialists begin the process of developing a database of research. At this stage, the personal data entered in the specialized statistical program, coded, and are multi-stage test. This technical activities that could prevent technical errors. For example, the incorrect coding can lead to loss of valuable data open-ended questions and data entry errors committed under the influence of human factors, may distort the data is properly carried out field stages. As a rule, carried out checks allow level technical input errors, and coding. Improper use of statistical analysis is more serious problem. Typically, marketing research data analysis is limited to simple tabulation – construction of linear and cross-sectional distributions of responses to the questionnaire. Of course, the use of elementary statistical procedures that do not require special skills and knowledge to mitigate the risk of data misinterpretation. But when used sophisticated statistical methods of cluster, regression, joint analysis, the risk of incorrect data increases. The output of the customer receives false information and is based on it are the same false conclusions. In conclusion, Marketing research – it highly intelligent activity. Luckily for professionals and amateurs to the chagrin of this case, the ability to conduct credible market research related to the wealth of knowledge in many scientific disciplines, and the mastery of this science takes many years. Nevertheless, today's growing market research there are more and more specialists with a distant view of reliability. Their formula – 'authentic' = 'true'. Their customers should be careful, taking for granted a 'truth'