On the dependable level of the negative survey

The negative survey is a new way to collect information, which could protect sensitive data and individual privacy. Generally, the results of negative survey need to be converted to positive survey results to acquire positive information. However, the traditional method to estimate positive surveys does not point out the dependable level of the negative survey. In this paper, the dependable level of the negative survey is analyzed by studying the confidence coefficient of the negative survey. And a greedy algorithm to calculate the shortest confidence intervals is proposed. Furthermore, the dependable level of multiple-choice negative survey is also analyzed.