Although conventional systems for social surveys offer various levels of privacy protection, patterns in the input data itself can accidentally lead to leakage of personal information. This is especially true in the case of small-scale social surveys. This research proposes a design for survey systems that can prevent this type of unintended information leakage. The basic design consists of a framework which analyzes the input data obtained to find elements that can lead to information leakage and a mechanism that automatically modifies the structure of the input data in database to correct this flaw.